wehub-resource-sync
1140 lines
70 KiB
Python
1140 lines
70 KiB
Python
"""Exhaustive parity sweep for the capability-ladder migration.
|
|
|
|
ModelCatalog.get_capabilities used to be a hardcoded per-provider prefix
|
|
ladder. That ladder's DATA now lives in catalog_overrides.toml corrections
|
|
rows and resolves through resolve_entry (only host-trust branches remain
|
|
code). This module freezes the pre-migration outputs as literals and asserts
|
|
the post-migration get_capabilities is byte-identical for EVERY registered
|
|
provider id crossed with a battery of model ids — each ladder prefix, a
|
|
miss, snapshot-known ids, and unknown ids — plus the openai host-guard
|
|
matrix (trusted host vs untrusted proxy) and the deepseek base-url sniff.
|
|
|
|
The expected literals were transcribed by running get_capabilities on the
|
|
UNMODIFIED tree (staging/provider-overhaul@43d6475c) via a one-off harness;
|
|
a sample is cross-checked against the pre-change ladder logic in
|
|
test_parity_legacy_capabilities_unchanged (test_catalog_layers.py). Do NOT
|
|
edit an expected tuple to make a test pass — a diff here is a real behavior
|
|
change in the migration.
|
|
|
|
Also freezes the two named provider sets moved into registry.py
|
|
(KEYLESS_PROVIDERS drives requires_api_key; LOCAL_RUNTIME_PROVIDERS drives
|
|
the catalog context-window heuristic) so unifying them — which would flip
|
|
requires_api_key("vllm") — fails loudly.
|
|
"""
|
|
|
|
from __future__ import annotations
|
|
|
|
from opensquilla.provider.model_catalog import DEFAULT_CONTEXT_WINDOW, ModelCatalog
|
|
from opensquilla.provider.registry import (
|
|
KEYLESS_PROVIDERS,
|
|
LOCAL_RUNTIME_PROVIDERS,
|
|
get_provider_spec,
|
|
list_provider_names,
|
|
list_provider_specs,
|
|
)
|
|
|
|
# (provider, model, base_url) ->
|
|
# (supports_reasoning, supports_tools, supports_vision, reasoning_format)
|
|
# Frozen from the pre-migration get_capabilities. supports_streaming is
|
|
# always True and asserted separately below.
|
|
_EXPECTED_CAPS: dict[tuple[str, str, str], tuple[bool, bool, bool, str]] = {
|
|
("aihubmix", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("aihubmix", "gpt-4o", ""): (False, True, True, "none"),
|
|
("aihubmix", "deepseek-r1", ""): (False, False, False, "none"),
|
|
("aihubmix", "glm-4.6", ""): (False, True, False, "none"),
|
|
("aihubmix", "qwen3-max", ""): (False, True, False, "none"),
|
|
("aihubmix", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("aihubmix", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("aihubmix", "kimi-latest", ""): (False, True, False, "none"),
|
|
("aihubmix", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("aihubmix", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("aihubmix", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("aihubmix", "gemini-2.5-flash", ""): (False, True, True, "none"),
|
|
("aihubmix", "gemini-3-pro-preview", ""): (False, True, True, "none"),
|
|
("aihubmix", "o3-mini", ""): (False, True, False, "none"),
|
|
("aihubmix", "gpt-5.5", ""): (False, True, True, "none"),
|
|
("anthropic", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("anthropic", "gpt-4o", ""): (False, True, False, "none"),
|
|
("anthropic", "deepseek-r1", ""): (False, True, False, "none"),
|
|
("anthropic", "glm-4.6", ""): (False, True, False, "none"),
|
|
("anthropic", "qwen3-max", ""): (False, True, False, "none"),
|
|
("anthropic", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("anthropic", "kimi-k2.5", ""): (False, True, False, "none"),
|
|
("anthropic", "kimi-latest", ""): (False, True, False, "none"),
|
|
("anthropic", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("anthropic", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("anthropic", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("anthropic", "gemini-2.5-flash", ""): (False, True, False, "none"),
|
|
("anthropic", "gemini-3-pro-preview", ""): (False, True, False, "none"),
|
|
("anthropic", "o3-mini", ""): (False, True, False, "none"),
|
|
("anthropic", "gpt-5.5", ""): (False, True, False, "none"),
|
|
("anthropic", "claude-opus-4-6", ""): (False, True, False, "none"),
|
|
("azure", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("azure", "gpt-4o", ""): (False, True, True, "none"),
|
|
("azure", "deepseek-r1", ""): (False, False, False, "none"),
|
|
("azure", "glm-4.6", ""): (False, True, False, "none"),
|
|
("azure", "qwen3-max", ""): (False, True, False, "none"),
|
|
("azure", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("azure", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("azure", "kimi-latest", ""): (False, True, False, "none"),
|
|
("azure", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("azure", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("azure", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("azure", "gemini-2.5-flash", ""): (False, True, True, "none"),
|
|
("azure", "gemini-3-pro-preview", ""): (False, True, True, "none"),
|
|
("azure", "o3-mini", ""): (False, True, False, "none"),
|
|
("azure", "gpt-5.5", ""): (False, True, True, "none"),
|
|
("bailian_coding", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("bailian_coding", "gpt-4o", ""): (False, True, True, "none"),
|
|
("bailian_coding", "deepseek-r1", ""): (False, True, False, "none"),
|
|
("bailian_coding", "glm-4.6", ""): (False, True, False, "none"),
|
|
("bailian_coding", "qwen3-max", ""): (False, True, False, "none"),
|
|
("bailian_coding", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("bailian_coding", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("bailian_coding", "kimi-latest", ""): (False, True, False, "none"),
|
|
("bailian_coding", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("bailian_coding", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("bailian_coding", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("bailian_coding", "gemini-2.5-flash", ""): (False, True, True, "none"),
|
|
("bailian_coding", "gemini-3-pro-preview", ""): (False, True, True, "none"),
|
|
("bailian_coding", "o3-mini", ""): (False, True, False, "none"),
|
|
("bailian_coding", "gpt-5.5", ""): (False, True, True, "none"),
|
|
("kimi_coding_anthropic", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("kimi_coding_anthropic", "gpt-4o", ""): (False, True, True, "none"),
|
|
("kimi_coding_anthropic", "deepseek-r1", ""): (False, False, False, "none"),
|
|
("kimi_coding_anthropic", "glm-4.6", ""): (False, True, False, "none"),
|
|
("kimi_coding_anthropic", "qwen3-max", ""): (False, True, False, "none"),
|
|
("kimi_coding_anthropic", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("kimi_coding_anthropic", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("kimi_coding_anthropic", "kimi-latest", ""): (False, True, False, "none"),
|
|
("kimi_coding_anthropic", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("kimi_coding_anthropic", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("kimi_coding_anthropic", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("kimi_coding_anthropic", "gemini-2.5-flash", ""): (False, True, True, "none"),
|
|
("kimi_coding_anthropic", "gemini-3-pro-preview", ""): (False, True, True, "none"),
|
|
("kimi_coding_anthropic", "o3-mini", ""): (False, True, False, "none"),
|
|
("kimi_coding_anthropic", "gpt-5.5", ""): (False, True, True, "none"),
|
|
("kimi_coding_anthropic", "kimi-for-coding", ""): (False, True, False, "none"),
|
|
("kimi_coding_openai", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("kimi_coding_openai", "gpt-4o", ""): (False, True, True, "none"),
|
|
("kimi_coding_openai", "deepseek-r1", ""): (False, False, False, "none"),
|
|
("kimi_coding_openai", "glm-4.6", ""): (False, True, False, "none"),
|
|
("kimi_coding_openai", "qwen3-max", ""): (False, True, False, "none"),
|
|
("kimi_coding_openai", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("kimi_coding_openai", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("kimi_coding_openai", "kimi-latest", ""): (False, True, False, "none"),
|
|
("kimi_coding_openai", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("kimi_coding_openai", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("kimi_coding_openai", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("kimi_coding_openai", "gemini-2.5-flash", ""): (False, True, True, "none"),
|
|
("kimi_coding_openai", "gemini-3-pro-preview", ""): (False, True, True, "none"),
|
|
("kimi_coding_openai", "o3-mini", ""): (False, True, False, "none"),
|
|
("kimi_coding_openai", "gpt-5.5", ""): (False, True, True, "none"),
|
|
("kimi_coding_openai", "kimi-for-coding", ""): (False, True, False, "none"),
|
|
("byteplus", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("byteplus", "gpt-4o", ""): (False, True, False, "none"),
|
|
("byteplus", "deepseek-r1", ""): (False, True, False, "none"),
|
|
("byteplus", "glm-4.6", ""): (False, True, False, "none"),
|
|
("byteplus", "qwen3-max", ""): (False, True, False, "none"),
|
|
("byteplus", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("byteplus", "kimi-k2.5", ""): (True, True, True, "volcengine"),
|
|
("byteplus", "kimi-latest", ""): (False, True, False, "none"),
|
|
("byteplus", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("byteplus", "seed-1-8-flash", ""): (True, True, True, "volcengine"),
|
|
("byteplus", "model-x-thinking", ""): (True, True, False, "volcengine"),
|
|
("byteplus", "gemini-2.5-flash", ""): (False, True, False, "none"),
|
|
("byteplus", "gemini-3-pro-preview", ""): (False, True, False, "none"),
|
|
("byteplus", "o3-mini", ""): (False, True, False, "none"),
|
|
("byteplus", "gpt-5.5", ""): (False, True, False, "none"),
|
|
("byteplus", "seed-1-6", ""): (True, True, True, "volcengine"),
|
|
("byteplus", "seed-1-6-250915", ""): (True, True, True, "volcengine"),
|
|
("byteplus", "seed-2-flash", ""): (True, True, True, "volcengine"),
|
|
("byteplus", "kimi-k2-250905", ""): (True, True, True, "volcengine"),
|
|
("byteplus", "kimi-k2.6", ""): (True, True, True, "volcengine"),
|
|
("byteplus", "kimi-k2-thinking", ""): (True, True, True, "volcengine"),
|
|
("byteplus", "skylark-pro", ""): (False, True, False, "none"),
|
|
("byteplus_coding_plan", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("byteplus_coding_plan", "gpt-4o", ""): (False, True, True, "none"),
|
|
("byteplus_coding_plan", "deepseek-r1", ""): (False, False, False, "none"),
|
|
("byteplus_coding_plan", "glm-4.6", ""): (False, True, False, "none"),
|
|
("byteplus_coding_plan", "qwen3-max", ""): (False, True, False, "none"),
|
|
("byteplus_coding_plan", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("byteplus_coding_plan", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("byteplus_coding_plan", "kimi-latest", ""): (False, True, False, "none"),
|
|
("byteplus_coding_plan", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("byteplus_coding_plan", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("byteplus_coding_plan", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("byteplus_coding_plan", "gemini-2.5-flash", ""): (False, True, True, "none"),
|
|
("byteplus_coding_plan", "gemini-3-pro-preview", ""): (False, True, True, "none"),
|
|
("byteplus_coding_plan", "o3-mini", ""): (False, True, False, "none"),
|
|
("byteplus_coding_plan", "gpt-5.5", ""): (False, True, True, "none"),
|
|
(
|
|
"byteplus_coding_plan_anthropic",
|
|
"totally-unknown-model-x1",
|
|
"",
|
|
): (False, True, False, "none"),
|
|
("custom", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("custom", "gpt-4o", ""): (False, True, True, "none"),
|
|
("custom", "deepseek-r1", ""): (False, False, False, "none"),
|
|
("custom", "glm-4.6", ""): (False, True, False, "none"),
|
|
("custom", "qwen3-max", ""): (False, True, False, "none"),
|
|
("custom", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("custom", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("custom", "kimi-latest", ""): (False, True, False, "none"),
|
|
("custom", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("custom", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("custom", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("custom", "gemini-2.5-flash", ""): (False, True, True, "none"),
|
|
("custom", "gemini-3-pro-preview", ""): (False, True, True, "none"),
|
|
("custom", "o3-mini", ""): (False, True, False, "none"),
|
|
("custom", "gpt-5.5", ""): (False, True, True, "none"),
|
|
("dashscope", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("dashscope", "gpt-4o", ""): (False, True, False, "none"),
|
|
("dashscope", "deepseek-r1", ""): (False, True, False, "none"),
|
|
("dashscope", "glm-4.6", ""): (False, True, False, "none"),
|
|
("dashscope", "qwen3-max", ""): (True, True, False, "dashscope"),
|
|
("dashscope", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("dashscope", "kimi-k2.5", ""): (False, True, False, "none"),
|
|
("dashscope", "kimi-latest", ""): (False, True, False, "none"),
|
|
("dashscope", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("dashscope", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("dashscope", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("dashscope", "gemini-2.5-flash", ""): (False, True, False, "none"),
|
|
("dashscope", "gemini-3-pro-preview", ""): (False, True, False, "none"),
|
|
("dashscope", "o3-mini", ""): (False, True, False, "none"),
|
|
("dashscope", "gpt-5.5", ""): (False, True, False, "none"),
|
|
("dashscope", "qwen3", ""): (True, True, False, "dashscope"),
|
|
("dashscope", "qwen3.5-vl-plus", ""): (True, True, True, "dashscope"),
|
|
("dashscope", "qwen3.6-omni", ""): (True, True, True, "dashscope"),
|
|
("dashscope", "qwen-plus-latest", ""): (True, True, False, "dashscope"),
|
|
("dashscope", "qwen-flash-2025", ""): (True, True, False, "dashscope"),
|
|
("dashscope", "qwen-turbo-x", ""): (True, True, False, "dashscope"),
|
|
("dashscope", "qwen-max-latest", ""): (True, True, False, "dashscope"),
|
|
("dashscope", "qwq-32b", ""): (True, True, False, "dashscope"),
|
|
("dashscope", "qwen-vl-ocr", ""): (False, True, True, "none"),
|
|
("dashscope", "qwen-vl-max", ""): (False, True, True, "none"),
|
|
("dashscope", "tongyi-intent-detect-v3", ""): (False, True, False, "none"),
|
|
("deepseek", "totally-unknown-model-x1", ""): (True, True, False, "deepseek"),
|
|
("deepseek", "gpt-4o", ""): (True, True, False, "deepseek"),
|
|
("deepseek", "deepseek-r1", ""): (True, True, False, "deepseek"),
|
|
("deepseek", "glm-4.6", ""): (True, True, False, "deepseek"),
|
|
("deepseek", "qwen3-max", ""): (True, True, False, "deepseek"),
|
|
("deepseek", "qwen2.5-14b-instruct", ""): (True, True, False, "deepseek"),
|
|
("deepseek", "kimi-k2.5", ""): (True, True, False, "deepseek"),
|
|
("deepseek", "kimi-latest", ""): (True, True, False, "deepseek"),
|
|
("deepseek", "doubao-seed-1-6-251015", ""): (True, True, False, "deepseek"),
|
|
("deepseek", "seed-1-8-flash", ""): (True, True, False, "deepseek"),
|
|
("deepseek", "model-x-thinking", ""): (True, True, False, "deepseek"),
|
|
("deepseek", "gemini-2.5-flash", ""): (True, True, False, "deepseek"),
|
|
("deepseek", "gemini-3-pro-preview", ""): (True, True, False, "deepseek"),
|
|
("deepseek", "o3-mini", ""): (True, True, False, "deepseek"),
|
|
("deepseek", "gpt-5.5", ""): (True, True, False, "deepseek"),
|
|
("deepseek", "deepseek-chat", ""): (True, True, False, "deepseek"),
|
|
("deepseek", "deepseek-reasoner", ""): (True, True, False, "deepseek"),
|
|
("deepseek", "deepseek-v4-flash", ""): (True, True, False, "deepseek"),
|
|
("gemini", "totally-unknown-model-x1", ""): (False, True, True, "none"),
|
|
("gemini", "gpt-4o", ""): (False, True, True, "none"),
|
|
("gemini", "deepseek-r1", ""): (False, True, True, "none"),
|
|
("gemini", "glm-4.6", ""): (False, True, True, "none"),
|
|
("gemini", "qwen3-max", ""): (False, True, True, "none"),
|
|
("gemini", "qwen2.5-14b-instruct", ""): (False, True, True, "none"),
|
|
("gemini", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("gemini", "kimi-latest", ""): (False, True, True, "none"),
|
|
("gemini", "doubao-seed-1-6-251015", ""): (False, True, True, "none"),
|
|
("gemini", "seed-1-8-flash", ""): (False, True, True, "none"),
|
|
("gemini", "model-x-thinking", ""): (False, True, True, "none"),
|
|
("gemini", "gemini-2.5-flash", ""): (True, True, True, "gemini"),
|
|
("gemini", "gemini-3-pro-preview", ""): (True, True, True, "gemini"),
|
|
("gemini", "o3-mini", ""): (False, True, True, "none"),
|
|
("gemini", "gpt-5.5", ""): (False, True, True, "none"),
|
|
("gemini", "gemini-2.5-pro", ""): (True, True, True, "gemini"),
|
|
("gemini", "gemini-2.0-flash", ""): (False, True, True, "none"),
|
|
("gemini", "gemini-2.5-flash-image", ""): (True, True, True, "gemini"),
|
|
("github_copilot", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("github_copilot", "gpt-4o", ""): (False, True, True, "none"),
|
|
("github_copilot", "deepseek-r1", ""): (False, False, False, "none"),
|
|
("github_copilot", "glm-4.6", ""): (False, True, False, "none"),
|
|
("github_copilot", "qwen3-max", ""): (False, True, False, "none"),
|
|
("github_copilot", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("github_copilot", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("github_copilot", "kimi-latest", ""): (False, True, False, "none"),
|
|
("github_copilot", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("github_copilot", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("github_copilot", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("github_copilot", "gemini-2.5-flash", ""): (False, True, True, "none"),
|
|
("github_copilot", "gemini-3-pro-preview", ""): (False, True, True, "none"),
|
|
("github_copilot", "o3-mini", ""): (False, True, False, "none"),
|
|
("github_copilot", "gpt-5.5", ""): (False, True, True, "none"),
|
|
("groq", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("groq", "gpt-4o", ""): (False, True, True, "none"),
|
|
("groq", "deepseek-r1", ""): (False, False, False, "none"),
|
|
("groq", "glm-4.6", ""): (False, True, False, "none"),
|
|
("groq", "qwen3-max", ""): (False, True, False, "none"),
|
|
("groq", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("groq", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("groq", "kimi-latest", ""): (False, True, False, "none"),
|
|
("groq", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("groq", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("groq", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("groq", "gemini-2.5-flash", ""): (False, True, True, "none"),
|
|
("groq", "gemini-3-pro-preview", ""): (False, True, True, "none"),
|
|
("groq", "o3-mini", ""): (False, True, False, "none"),
|
|
("groq", "gpt-5.5", ""): (False, True, True, "none"),
|
|
("litellm_proxy", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("litellm_proxy", "gpt-4o", ""): (False, True, True, "none"),
|
|
("litellm_proxy", "deepseek-r1", ""): (False, False, False, "none"),
|
|
("litellm_proxy", "glm-4.6", ""): (False, True, False, "none"),
|
|
("litellm_proxy", "qwen3-max", ""): (False, True, False, "none"),
|
|
("litellm_proxy", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("litellm_proxy", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("litellm_proxy", "kimi-latest", ""): (False, True, False, "none"),
|
|
("litellm_proxy", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("litellm_proxy", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("litellm_proxy", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("litellm_proxy", "gemini-2.5-flash", ""): (False, True, True, "none"),
|
|
("litellm_proxy", "gemini-3-pro-preview", ""): (False, True, True, "none"),
|
|
("litellm_proxy", "o3-mini", ""): (False, True, False, "none"),
|
|
("litellm_proxy", "gpt-5.5", ""): (False, True, True, "none"),
|
|
("lm_studio", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("lm_studio", "gpt-4o", ""): (False, True, True, "none"),
|
|
("lm_studio", "deepseek-r1", ""): (False, False, False, "none"),
|
|
("lm_studio", "glm-4.6", ""): (False, True, False, "none"),
|
|
("lm_studio", "qwen3-max", ""): (False, True, False, "none"),
|
|
("lm_studio", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("lm_studio", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("lm_studio", "kimi-latest", ""): (False, True, False, "none"),
|
|
("lm_studio", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("lm_studio", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("lm_studio", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("lm_studio", "gemini-2.5-flash", ""): (False, True, True, "none"),
|
|
("lm_studio", "gemini-3-pro-preview", ""): (False, True, True, "none"),
|
|
("lm_studio", "o3-mini", ""): (False, True, False, "none"),
|
|
("lm_studio", "gpt-5.5", ""): (False, True, True, "none"),
|
|
("minimax", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("minimax", "gpt-4o", ""): (False, True, True, "none"),
|
|
("minimax", "deepseek-r1", ""): (False, False, False, "none"),
|
|
("minimax", "glm-4.6", ""): (False, True, False, "none"),
|
|
("minimax", "qwen3-max", ""): (False, True, False, "none"),
|
|
("minimax", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("minimax", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("minimax", "kimi-latest", ""): (False, True, False, "none"),
|
|
("minimax", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("minimax", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("minimax", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("minimax", "gemini-2.5-flash", ""): (False, True, True, "none"),
|
|
("minimax", "gemini-3-pro-preview", ""): (False, True, True, "none"),
|
|
("minimax", "o3-mini", ""): (False, True, False, "none"),
|
|
("minimax", "gpt-5.5", ""): (False, True, True, "none"),
|
|
("minimax_cn", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("minimax_cn", "gpt-4o", ""): (False, True, True, "none"),
|
|
("minimax_cn", "deepseek-r1", ""): (False, False, False, "none"),
|
|
("minimax_cn", "glm-4.6", ""): (False, True, False, "none"),
|
|
("minimax_cn", "qwen3-max", ""): (False, True, False, "none"),
|
|
("minimax_cn", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("minimax_cn", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("minimax_cn", "kimi-latest", ""): (False, True, False, "none"),
|
|
("minimax_cn", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("minimax_cn", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("minimax_cn", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("minimax_cn", "gemini-2.5-flash", ""): (False, True, True, "none"),
|
|
("minimax_cn", "gemini-3-pro-preview", ""): (False, True, True, "none"),
|
|
("minimax_cn", "o3-mini", ""): (False, True, False, "none"),
|
|
("minimax_cn", "gpt-5.5", ""): (False, True, True, "none"),
|
|
("minimax_global", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("minimax_global", "gpt-4o", ""): (False, True, True, "none"),
|
|
("minimax_global", "deepseek-r1", ""): (False, False, False, "none"),
|
|
("minimax_global", "glm-4.6", ""): (False, True, False, "none"),
|
|
("minimax_global", "qwen3-max", ""): (False, True, False, "none"),
|
|
("minimax_global", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("minimax_global", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("minimax_global", "kimi-latest", ""): (False, True, False, "none"),
|
|
("minimax_global", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("minimax_global", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("minimax_global", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("minimax_global", "gemini-2.5-flash", ""): (False, True, True, "none"),
|
|
("minimax_global", "gemini-3-pro-preview", ""): (False, True, True, "none"),
|
|
("minimax_global", "o3-mini", ""): (False, True, False, "none"),
|
|
("minimax_global", "gpt-5.5", ""): (False, True, True, "none"),
|
|
("minimax_openai", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("minimax_openai", "gpt-4o", ""): (False, True, True, "none"),
|
|
("minimax_openai", "deepseek-r1", ""): (False, False, False, "none"),
|
|
("minimax_openai", "glm-4.6", ""): (False, True, False, "none"),
|
|
("minimax_openai", "qwen3-max", ""): (False, True, False, "none"),
|
|
("minimax_openai", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("minimax_openai", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("minimax_openai", "kimi-latest", ""): (False, True, False, "none"),
|
|
("minimax_openai", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("minimax_openai", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("minimax_openai", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("minimax_openai", "gemini-2.5-flash", ""): (False, True, True, "none"),
|
|
("minimax_openai", "gemini-3-pro-preview", ""): (False, True, True, "none"),
|
|
("minimax_openai", "o3-mini", ""): (False, True, False, "none"),
|
|
("minimax_openai", "gpt-5.5", ""): (False, True, True, "none"),
|
|
("minimax_coding_anthropic", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("minimax_coding_anthropic", "gpt-4o", ""): (False, True, True, "none"),
|
|
("minimax_coding_anthropic", "deepseek-r1", ""): (False, False, False, "none"),
|
|
("minimax_coding_anthropic", "glm-4.6", ""): (False, True, False, "none"),
|
|
("minimax_coding_anthropic", "qwen3-max", ""): (False, True, False, "none"),
|
|
("minimax_coding_anthropic", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("minimax_coding_anthropic", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("minimax_coding_anthropic", "kimi-latest", ""): (False, True, False, "none"),
|
|
("minimax_coding_anthropic", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("minimax_coding_anthropic", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("minimax_coding_anthropic", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("minimax_coding_anthropic", "gemini-2.5-flash", ""): (False, True, True, "none"),
|
|
("minimax_coding_anthropic", "gemini-3-pro-preview", ""): (False, True, True, "none"),
|
|
("minimax_coding_anthropic", "o3-mini", ""): (False, True, False, "none"),
|
|
("minimax_coding_anthropic", "gpt-5.5", ""): (False, True, True, "none"),
|
|
("minimax_coding_anthropic", "MiniMax-M2.7", ""): (False, True, False, "none"),
|
|
("minimax_coding_anthropic", "MiniMax-M3", ""): (False, True, False, "none"),
|
|
("minimax_coding_openai", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("minimax_coding_openai", "gpt-4o", ""): (False, True, True, "none"),
|
|
("minimax_coding_openai", "deepseek-r1", ""): (False, False, False, "none"),
|
|
("minimax_coding_openai", "glm-4.6", ""): (False, True, False, "none"),
|
|
("minimax_coding_openai", "qwen3-max", ""): (False, True, False, "none"),
|
|
("minimax_coding_openai", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("minimax_coding_openai", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("minimax_coding_openai", "kimi-latest", ""): (False, True, False, "none"),
|
|
("minimax_coding_openai", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("minimax_coding_openai", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("minimax_coding_openai", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("minimax_coding_openai", "gemini-2.5-flash", ""): (False, True, True, "none"),
|
|
("minimax_coding_openai", "gemini-3-pro-preview", ""): (False, True, True, "none"),
|
|
("minimax_coding_openai", "o3-mini", ""): (False, True, False, "none"),
|
|
("minimax_coding_openai", "gpt-5.5", ""): (False, True, True, "none"),
|
|
("minimax_coding_openai", "MiniMax-M2.7", ""): (False, True, False, "none"),
|
|
("minimax_coding_openai", "MiniMax-M3", ""): (False, True, False, "none"),
|
|
("mimo_anthropic", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("mimo_anthropic", "gpt-4o", ""): (False, True, True, "none"),
|
|
("mimo_anthropic", "deepseek-r1", ""): (False, False, False, "none"),
|
|
("mimo_anthropic", "glm-4.6", ""): (False, True, False, "none"),
|
|
("mimo_anthropic", "qwen3-max", ""): (False, True, False, "none"),
|
|
("mimo_anthropic", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("mimo_anthropic", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("mimo_anthropic", "kimi-latest", ""): (False, True, False, "none"),
|
|
("mimo_anthropic", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("mimo_anthropic", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("mimo_anthropic", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("mimo_anthropic", "gemini-2.5-flash", ""): (False, True, True, "none"),
|
|
("mimo_anthropic", "gemini-3-pro-preview", ""): (False, True, True, "none"),
|
|
("mimo_anthropic", "o3-mini", ""): (False, True, False, "none"),
|
|
("mimo_anthropic", "gpt-5.5", ""): (False, True, True, "none"),
|
|
("mimo_anthropic", "mimo-v2.5", ""): (False, True, False, "none"),
|
|
("mimo_anthropic", "mimo-v2.5-pro", ""): (False, True, False, "none"),
|
|
("mimo_openai", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("mimo_openai", "gpt-4o", ""): (False, True, True, "none"),
|
|
("mimo_openai", "deepseek-r1", ""): (False, False, False, "none"),
|
|
("mimo_openai", "glm-4.6", ""): (False, True, False, "none"),
|
|
("mimo_openai", "qwen3-max", ""): (False, True, False, "none"),
|
|
("mimo_openai", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("mimo_openai", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("mimo_openai", "kimi-latest", ""): (False, True, False, "none"),
|
|
("mimo_openai", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("mimo_openai", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("mimo_openai", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("mimo_openai", "gemini-2.5-flash", ""): (False, True, True, "none"),
|
|
("mimo_openai", "gemini-3-pro-preview", ""): (False, True, True, "none"),
|
|
("mimo_openai", "o3-mini", ""): (False, True, False, "none"),
|
|
("mimo_openai", "gpt-5.5", ""): (False, True, True, "none"),
|
|
("mimo_openai", "mimo-v2.5", ""): (False, True, False, "none"),
|
|
("mimo_openai", "mimo-v2.5-pro", ""): (False, True, False, "none"),
|
|
("mistral", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("mistral", "gpt-4o", ""): (False, True, True, "none"),
|
|
("mistral", "deepseek-r1", ""): (False, False, False, "none"),
|
|
("mistral", "glm-4.6", ""): (False, True, False, "none"),
|
|
("mistral", "qwen3-max", ""): (False, True, False, "none"),
|
|
("mistral", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("mistral", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("mistral", "kimi-latest", ""): (False, True, False, "none"),
|
|
("mistral", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("mistral", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("mistral", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("mistral", "gemini-2.5-flash", ""): (False, True, True, "none"),
|
|
("mistral", "gemini-3-pro-preview", ""): (False, True, True, "none"),
|
|
("mistral", "o3-mini", ""): (False, True, False, "none"),
|
|
("mistral", "gpt-5.5", ""): (False, True, True, "none"),
|
|
("moonshot", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("moonshot", "gpt-4o", ""): (False, True, False, "none"),
|
|
("moonshot", "deepseek-r1", ""): (False, True, False, "none"),
|
|
("moonshot", "glm-4.6", ""): (False, True, False, "none"),
|
|
("moonshot", "qwen3-max", ""): (False, True, False, "none"),
|
|
("moonshot", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("moonshot", "kimi-k2.5", ""): (True, True, True, "moonshot"),
|
|
("moonshot", "kimi-latest", ""): (False, True, False, "none"),
|
|
("moonshot", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("moonshot", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("moonshot", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("moonshot", "gemini-2.5-flash", ""): (False, True, False, "none"),
|
|
("moonshot", "gemini-3-pro-preview", ""): (False, True, False, "none"),
|
|
("moonshot", "o3-mini", ""): (False, True, False, "none"),
|
|
("moonshot", "gpt-5.5", ""): (False, True, False, "none"),
|
|
("moonshot", "kimi-k2.6-preview", ""): (True, True, True, "moonshot"),
|
|
("moonshot", "kimi-k2-thinking", ""): (True, True, False, "moonshot"),
|
|
("moonshot", "kimi-k2-thinking-turbo", ""): (True, True, False, "moonshot"),
|
|
("moonshot", "kimi-k2-0905-preview", ""): (False, True, False, "none"),
|
|
("moonshot", "moonshot-v1-8k", ""): (False, True, False, "none"),
|
|
("moonshot", "moonshot-v1-128k", ""): (False, True, False, "none"),
|
|
("ollama", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("ollama", "gpt-4o", ""): (False, True, False, "none"),
|
|
("ollama", "deepseek-r1", ""): (False, True, False, "none"),
|
|
("ollama", "glm-4.6", ""): (False, True, False, "none"),
|
|
("ollama", "qwen3-max", ""): (False, True, False, "none"),
|
|
("ollama", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("ollama", "kimi-k2.5", ""): (False, True, False, "none"),
|
|
("ollama", "kimi-latest", ""): (False, True, False, "none"),
|
|
("ollama", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("ollama", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("ollama", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("ollama", "gemini-2.5-flash", ""): (False, True, False, "none"),
|
|
("ollama", "gemini-3-pro-preview", ""): (False, True, False, "none"),
|
|
("ollama", "o3-mini", ""): (False, True, False, "none"),
|
|
("ollama", "gpt-5.5", ""): (False, True, False, "none"),
|
|
("ollama", "llama3.2:3b", ""): (False, True, False, "none"),
|
|
("openai", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("openai", "gpt-4o", ""): (False, True, True, "none"),
|
|
("openai", "deepseek-r1", ""): (False, False, False, "none"),
|
|
("openai", "glm-4.6", ""): (False, True, False, "none"),
|
|
("openai", "qwen3-max", ""): (False, True, False, "none"),
|
|
("openai", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("openai", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("openai", "kimi-latest", ""): (False, True, False, "none"),
|
|
("openai", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("openai", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("openai", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("openai", "gemini-2.5-flash", ""): (False, True, True, "none"),
|
|
("openai", "gemini-3-pro-preview", ""): (False, True, True, "none"),
|
|
("openai", "o3-mini", ""): (False, True, False, "none"),
|
|
("openai", "gpt-5.5", ""): (False, True, True, "none"),
|
|
("openai", "gpt-3.5-turbo", ""): (False, False, False, "none"),
|
|
("openai_codex", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("openai_codex", "gpt-4o", ""): (False, True, True, "none"),
|
|
("openai_codex", "deepseek-r1", ""): (False, False, False, "none"),
|
|
("openai_codex", "glm-4.6", ""): (False, True, False, "none"),
|
|
("openai_codex", "qwen3-max", ""): (False, True, False, "none"),
|
|
("openai_codex", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("openai_codex", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("openai_codex", "kimi-latest", ""): (False, True, False, "none"),
|
|
("openai_codex", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("openai_codex", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("openai_codex", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("openai_codex", "gemini-2.5-flash", ""): (False, True, True, "none"),
|
|
("openai_codex", "gemini-3-pro-preview", ""): (False, True, True, "none"),
|
|
("openai_codex", "o3-mini", ""): (False, True, False, "none"),
|
|
("openai_codex", "gpt-5.5", ""): (False, True, True, "none"),
|
|
("openai_responses", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("openai_responses", "gpt-4o", ""): (False, True, True, "none"),
|
|
("openai_responses", "deepseek-r1", ""): (False, False, False, "none"),
|
|
("openai_responses", "glm-4.6", ""): (False, True, False, "none"),
|
|
("openai_responses", "qwen3-max", ""): (False, True, False, "none"),
|
|
("openai_responses", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("openai_responses", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("openai_responses", "kimi-latest", ""): (False, True, False, "none"),
|
|
("openai_responses", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("openai_responses", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("openai_responses", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("openai_responses", "gemini-2.5-flash", ""): (False, True, True, "none"),
|
|
("openai_responses", "gemini-3-pro-preview", ""): (False, True, True, "none"),
|
|
("openai_responses", "o3-mini", ""): (False, True, False, "none"),
|
|
("openai_responses", "gpt-5.5", ""): (False, True, True, "none"),
|
|
("openrouter", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("openrouter", "gpt-4o", ""): (False, True, True, "none"),
|
|
("openrouter", "deepseek-r1", ""): (False, False, False, "none"),
|
|
("openrouter", "glm-4.6", ""): (False, True, False, "none"),
|
|
("openrouter", "qwen3-max", ""): (False, True, False, "none"),
|
|
("openrouter", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("openrouter", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("openrouter", "kimi-latest", ""): (False, True, False, "none"),
|
|
("openrouter", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("openrouter", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("openrouter", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("openrouter", "gemini-2.5-flash", ""): (False, True, True, "none"),
|
|
("openrouter", "gemini-3-pro-preview", ""): (False, True, True, "none"),
|
|
("openrouter", "o3-mini", ""): (False, True, False, "none"),
|
|
("openrouter", "gpt-5.5", ""): (False, True, True, "none"),
|
|
# Snapshot-known reasoning model without tools. Originally
|
|
# aion-labs/aion-1.0; models.dev dropped that id (2026-07-08 refresh), so
|
|
# the same scenario is pinned on another snapshot row with the same shape
|
|
# (reasoning=true suppressed without a dialect, tools=false vendored).
|
|
("openrouter", "tencent/hunyuan-a13b-instruct", ""): (False, False, False, "none"),
|
|
("openrouter", "amazon/nova-2-lite-v1", ""): (False, True, True, "none"),
|
|
("openrouter", "z-ai/glm-5", ""): (False, True, False, "none"),
|
|
("ovms", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("ovms", "gpt-4o", ""): (False, True, True, "none"),
|
|
("ovms", "deepseek-r1", ""): (False, False, False, "none"),
|
|
("ovms", "glm-4.6", ""): (False, True, False, "none"),
|
|
("ovms", "qwen3-max", ""): (False, True, False, "none"),
|
|
("ovms", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("ovms", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("ovms", "kimi-latest", ""): (False, True, False, "none"),
|
|
("ovms", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("ovms", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("ovms", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("ovms", "gemini-2.5-flash", ""): (False, True, True, "none"),
|
|
("ovms", "gemini-3-pro-preview", ""): (False, True, True, "none"),
|
|
("ovms", "o3-mini", ""): (False, True, False, "none"),
|
|
("ovms", "gpt-5.5", ""): (False, True, True, "none"),
|
|
("qianfan", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("qianfan", "gpt-4o", ""): (False, True, True, "none"),
|
|
("qianfan", "deepseek-r1", ""): (False, False, False, "none"),
|
|
("qianfan", "glm-4.6", ""): (False, True, False, "none"),
|
|
("qianfan", "qwen3-max", ""): (False, True, False, "none"),
|
|
("qianfan", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("qianfan", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("qianfan", "kimi-latest", ""): (False, True, False, "none"),
|
|
("qianfan", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("qianfan", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("qianfan", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("qianfan", "gemini-2.5-flash", ""): (False, True, True, "none"),
|
|
("qianfan", "gemini-3-pro-preview", ""): (False, True, True, "none"),
|
|
("qianfan", "o3-mini", ""): (False, True, False, "none"),
|
|
("qianfan", "gpt-5.5", ""): (False, True, True, "none"),
|
|
("siliconflow", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("siliconflow", "gpt-4o", ""): (False, True, True, "none"),
|
|
("siliconflow", "deepseek-r1", ""): (False, False, False, "none"),
|
|
("siliconflow", "glm-4.6", ""): (False, True, False, "none"),
|
|
("siliconflow", "qwen3-max", ""): (False, True, False, "none"),
|
|
("siliconflow", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("siliconflow", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("siliconflow", "kimi-latest", ""): (False, True, False, "none"),
|
|
("siliconflow", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("siliconflow", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("siliconflow", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("siliconflow", "gemini-2.5-flash", ""): (False, True, True, "none"),
|
|
("siliconflow", "gemini-3-pro-preview", ""): (False, True, True, "none"),
|
|
("siliconflow", "o3-mini", ""): (False, True, False, "none"),
|
|
("siliconflow", "gpt-5.5", ""): (False, True, True, "none"),
|
|
("tencent_token_plan", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("tencent_token_plan", "gpt-4o", ""): (False, True, False, "none"),
|
|
("tencent_token_plan", "deepseek-r1", ""): (False, True, False, "none"),
|
|
("tencent_token_plan", "glm-4.6", ""): (False, True, False, "none"),
|
|
("tencent_token_plan", "qwen3-max", ""): (False, True, False, "none"),
|
|
("tencent_token_plan", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("tencent_token_plan", "kimi-k2.5", ""): (False, True, False, "none"),
|
|
("tencent_token_plan", "kimi-latest", ""): (False, True, False, "none"),
|
|
("tencent_token_plan", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("tencent_token_plan", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("tencent_token_plan", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("tencent_token_plan", "gemini-2.5-flash", ""): (False, True, False, "none"),
|
|
("tencent_token_plan", "gemini-3-pro-preview", ""): (False, True, False, "none"),
|
|
("tencent_token_plan", "o3-mini", ""): (False, True, False, "none"),
|
|
("tencent_token_plan", "gpt-5.5", ""): (False, True, False, "none"),
|
|
("tencent_token_plan", "hy3", ""): (True, True, False, "tencent_tokenhub"),
|
|
("tencent_token_plan", "hy3-preview", ""): (True, True, False, "tencent_tokenhub"),
|
|
("tencent_token_plan_anthropic", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("tencent_token_plan_anthropic", "gpt-4o", ""): (False, True, True, "none"),
|
|
("tencent_token_plan_anthropic", "deepseek-r1", ""): (False, False, False, "none"),
|
|
("tencent_token_plan_anthropic", "glm-4.6", ""): (False, True, False, "none"),
|
|
("tencent_token_plan_anthropic", "qwen3-max", ""): (False, True, False, "none"),
|
|
("tencent_token_plan_anthropic", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("tencent_token_plan_anthropic", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("tencent_token_plan_anthropic", "kimi-latest", ""): (False, True, False, "none"),
|
|
("tencent_token_plan_anthropic", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("tencent_token_plan_anthropic", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("tencent_token_plan_anthropic", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("tencent_token_plan_anthropic", "gemini-2.5-flash", ""): (False, True, True, "none"),
|
|
("tencent_token_plan_anthropic", "gemini-3-pro-preview", ""): (False, True, True, "none"),
|
|
("tencent_token_plan_anthropic", "o3-mini", ""): (False, True, False, "none"),
|
|
("tencent_token_plan_anthropic", "gpt-5.5", ""): (False, True, True, "none"),
|
|
("tencent_token_plan_anthropic", "hy3", ""): (False, True, False, "none"),
|
|
("tencent_token_plan_anthropic", "hy3-preview", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub", "gpt-4o", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub", "deepseek-r1", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub", "glm-4.6", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub", "qwen3-max", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub", "kimi-k2.5", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub", "kimi-latest", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub", "gemini-2.5-flash", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub", "gemini-3-pro-preview", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub", "o3-mini", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub", "gpt-5.5", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub", "hy3", ""): (True, True, False, "tencent_tokenhub"),
|
|
("tencent_tokenhub", "hy3-preview", ""): (True, True, False, "tencent_tokenhub"),
|
|
("tencent_tokenhub_anthropic", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub_anthropic", "gpt-4o", ""): (False, True, True, "none"),
|
|
("tencent_tokenhub_anthropic", "deepseek-r1", ""): (False, False, False, "none"),
|
|
("tencent_tokenhub_anthropic", "glm-4.6", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub_anthropic", "qwen3-max", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub_anthropic", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub_anthropic", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("tencent_tokenhub_anthropic", "kimi-latest", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub_anthropic", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub_anthropic", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub_anthropic", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub_anthropic", "gemini-2.5-flash", ""): (False, True, True, "none"),
|
|
("tencent_tokenhub_anthropic", "gemini-3-pro-preview", ""): (False, True, True, "none"),
|
|
("tencent_tokenhub_anthropic", "o3-mini", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub_anthropic", "gpt-5.5", ""): (False, True, True, "none"),
|
|
("tencent_tokenhub_anthropic", "hy3", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub_anthropic", "hy3-preview", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub_intl", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub_intl", "gpt-4o", ""): (False, True, True, "none"),
|
|
("tencent_tokenhub_intl", "deepseek-r1", ""): (False, False, False, "none"),
|
|
("tencent_tokenhub_intl", "glm-4.6", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub_intl", "qwen3-max", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub_intl", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub_intl", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("tencent_tokenhub_intl", "kimi-latest", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub_intl", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub_intl", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub_intl", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub_intl", "gemini-2.5-flash", ""): (False, True, True, "none"),
|
|
("tencent_tokenhub_intl", "gemini-3-pro-preview", ""): (False, True, True, "none"),
|
|
("tencent_tokenhub_intl", "o3-mini", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub_intl", "gpt-5.5", ""): (False, True, True, "none"),
|
|
("tencent_tokenhub_intl", "hy3", ""): (False, True, False, "none"),
|
|
("tencent_tokenhub_intl", "hy3-preview", ""): (False, True, False, "none"),
|
|
# tokenrhythm: the [tokenrhythm.*] corrections rows pin
|
|
# reasoning_format="none" (the relay streams reasoning_content on its
|
|
# own but rejects thinking toggles), so effective supports_reasoning is
|
|
# False everywhere; vision only on the live-verified kimi rows.
|
|
("tokenrhythm", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("tokenrhythm", "gpt-4o", ""): (False, True, False, "none"),
|
|
("tokenrhythm", "deepseek-r1", ""): (False, True, False, "none"),
|
|
("tokenrhythm", "glm-4.6", ""): (False, True, False, "none"),
|
|
("tokenrhythm", "qwen3-max", ""): (False, True, False, "none"),
|
|
("tokenrhythm", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("tokenrhythm", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("tokenrhythm", "kimi-latest", ""): (False, True, False, "none"),
|
|
("tokenrhythm", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("tokenrhythm", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("tokenrhythm", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("tokenrhythm", "gemini-2.5-flash", ""): (False, True, False, "none"),
|
|
("tokenrhythm", "gemini-3-pro-preview", ""): (False, True, False, "none"),
|
|
("tokenrhythm", "o3-mini", ""): (False, True, False, "none"),
|
|
("tokenrhythm", "gpt-5.5", ""): (False, True, False, "none"),
|
|
("tokenrhythm", "deepseek-v4-flash", ""): (False, True, False, "none"),
|
|
("tokenrhythm", "kimi-k2.7-code", ""): (False, True, True, "none"),
|
|
("vllm", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("vllm", "gpt-4o", ""): (False, True, True, "none"),
|
|
("vllm", "deepseek-r1", ""): (False, False, False, "none"),
|
|
("vllm", "glm-4.6", ""): (False, True, False, "none"),
|
|
("vllm", "qwen3-max", ""): (False, True, False, "none"),
|
|
("vllm", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("vllm", "kimi-k2.5", ""): (False, True, True, "none"),
|
|
("vllm", "kimi-latest", ""): (False, True, False, "none"),
|
|
("vllm", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("vllm", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("vllm", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("vllm", "gemini-2.5-flash", ""): (False, True, True, "none"),
|
|
("vllm", "gemini-3-pro-preview", ""): (False, True, True, "none"),
|
|
("vllm", "o3-mini", ""): (False, True, False, "none"),
|
|
("vllm", "gpt-5.5", ""): (False, True, True, "none"),
|
|
("volcengine", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("volcengine", "gpt-4o", ""): (False, True, False, "none"),
|
|
("volcengine", "deepseek-r1", ""): (False, True, False, "none"),
|
|
("volcengine", "glm-4.6", ""): (False, True, False, "none"),
|
|
("volcengine", "qwen3-max", ""): (False, True, False, "none"),
|
|
("volcengine", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("volcengine", "kimi-k2.5", ""): (False, True, False, "none"),
|
|
("volcengine", "kimi-latest", ""): (False, True, False, "none"),
|
|
("volcengine", "doubao-seed-1-6-251015", ""): (True, True, True, "volcengine"),
|
|
("volcengine", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("volcengine", "model-x-thinking", ""): (True, True, False, "volcengine"),
|
|
("volcengine", "gemini-2.5-flash", ""): (False, True, False, "none"),
|
|
("volcengine", "gemini-3-pro-preview", ""): (False, True, False, "none"),
|
|
("volcengine", "o3-mini", ""): (False, True, False, "none"),
|
|
("volcengine", "gpt-5.5", ""): (False, True, False, "none"),
|
|
("volcengine", "doubao-seed-1-8-250915", ""): (True, True, True, "volcengine"),
|
|
("volcengine", "doubao-seed-2-pro", ""): (True, True, True, "volcengine"),
|
|
("volcengine", "doubao-1-5-thinking-pro-250415", ""): (True, True, False, "volcengine"),
|
|
("volcengine", "doubao-pro-32k", ""): (False, True, False, "none"),
|
|
("volcengine", "kimi-k2-thinking", ""): (True, True, False, "volcengine"),
|
|
("volcengine_coding_plan", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("volcengine_coding_plan", "gpt-4o", ""): (False, True, False, "none"),
|
|
("volcengine_coding_plan", "deepseek-r1", ""): (False, True, False, "none"),
|
|
("volcengine_coding_plan", "glm-4.6", ""): (False, True, False, "none"),
|
|
("volcengine_coding_plan", "qwen3-max", ""): (False, True, False, "none"),
|
|
("volcengine_coding_plan", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("volcengine_coding_plan", "kimi-k2.5", ""): (False, True, False, "none"),
|
|
("volcengine_coding_plan", "kimi-latest", ""): (False, True, False, "none"),
|
|
("volcengine_coding_plan", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("volcengine_coding_plan", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("volcengine_coding_plan", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("volcengine_coding_plan", "gemini-2.5-flash", ""): (False, True, False, "none"),
|
|
("volcengine_coding_plan", "gemini-3-pro-preview", ""): (False, True, False, "none"),
|
|
("volcengine_coding_plan", "o3-mini", ""): (False, True, False, "none"),
|
|
("volcengine_coding_plan", "gpt-5.5", ""): (False, True, False, "none"),
|
|
(
|
|
"volcengine_coding_plan_anthropic",
|
|
"totally-unknown-model-x1",
|
|
"",
|
|
): (False, True, False, "none"),
|
|
("volcengine_coding_plan", "doubao-seed-2.0-lite", ""): (True, True, True, "volcengine"),
|
|
("volcengine_coding_plan", "doubao-seed-2.0-pro", ""): (True, True, True, "volcengine"),
|
|
(
|
|
"volcengine_coding_plan",
|
|
"doubao-seed-2.0-code",
|
|
"",
|
|
): (True, True, True, "volcengine"),
|
|
("zhipu", "totally-unknown-model-x1", ""): (False, True, False, "none"),
|
|
("zhipu", "gpt-4o", ""): (False, True, False, "none"),
|
|
("zhipu", "deepseek-r1", ""): (False, True, False, "none"),
|
|
("zhipu", "glm-4.6", ""): (False, True, False, "none"),
|
|
("zhipu", "qwen3-max", ""): (False, True, False, "none"),
|
|
("zhipu", "qwen2.5-14b-instruct", ""): (False, True, False, "none"),
|
|
("zhipu", "kimi-k2.5", ""): (False, True, False, "none"),
|
|
("zhipu", "kimi-latest", ""): (False, True, False, "none"),
|
|
("zhipu", "doubao-seed-1-6-251015", ""): (False, True, False, "none"),
|
|
("zhipu", "seed-1-8-flash", ""): (False, True, False, "none"),
|
|
("zhipu", "model-x-thinking", ""): (False, True, False, "none"),
|
|
("zhipu", "gemini-2.5-flash", ""): (False, True, False, "none"),
|
|
("zhipu", "gemini-3-pro-preview", ""): (False, True, False, "none"),
|
|
("zhipu", "o3-mini", ""): (False, True, False, "none"),
|
|
("zhipu", "gpt-5.5", ""): (False, True, False, "none"),
|
|
("zhipu", "glm-4.5-air", ""): (True, True, False, "zai"),
|
|
("zhipu", "glm-4.5v", ""): (True, True, False, "zai"),
|
|
("zhipu", "glm-4.7-flashx", ""): (True, True, False, "zai"),
|
|
("zhipu", "glm-5", ""): (True, True, False, "zai"),
|
|
("zhipu", "glm-5.2", ""): (True, True, False, "zai"),
|
|
("zhipu", "glm-4", ""): (False, True, False, "none"),
|
|
("openai", "gpt-5.5", "https://api.openai.com/v1"): (True, True, False, "openai"),
|
|
("openai", "gpt-5.4", "https://api.openai.com/v1"): (True, True, False, "openai"),
|
|
("openai", "gpt-5", "https://api.openai.com/v1"): (True, True, False, "openai"),
|
|
("openai", "o1", "https://api.openai.com/v1"): (True, True, False, "openai"),
|
|
("openai", "o1-preview", "https://api.openai.com/v1"): (True, True, False, "openai"),
|
|
("openai", "o3", "https://api.openai.com/v1"): (True, True, False, "openai"),
|
|
("openai", "o3-mini", "https://api.openai.com/v1"): (True, True, False, "openai"),
|
|
("openai", "o4-mini", "https://api.openai.com/v1"): (True, True, False, "openai"),
|
|
("openai", "gpt-4o", "https://api.openai.com/v1"): (False, True, True, "none"),
|
|
("openai", "gpt-4.1", "https://api.openai.com/v1"): (False, True, True, "none"),
|
|
("openai", "gpt-3.5-turbo", "https://api.openai.com/v1"): (False, False, False, "none"),
|
|
("openai", "omni-model", "https://api.openai.com/v1"): (False, True, False, "none"),
|
|
("openai", "gpt-5.5", "https://proxy.example/v1"): (False, True, True, "none"),
|
|
("openai", "gpt-5.4", "https://proxy.example/v1"): (False, True, True, "none"),
|
|
("openai", "gpt-5", "https://proxy.example/v1"): (False, True, True, "none"),
|
|
("openai", "o1", "https://proxy.example/v1"): (False, True, True, "none"),
|
|
("openai", "o1-preview", "https://proxy.example/v1"): (False, True, False, "none"),
|
|
("openai", "o3", "https://proxy.example/v1"): (False, True, True, "none"),
|
|
("openai", "o3-mini", "https://proxy.example/v1"): (False, True, False, "none"),
|
|
("openai", "o4-mini", "https://proxy.example/v1"): (False, True, True, "none"),
|
|
("openai", "gpt-4o", "https://proxy.example/v1"): (False, True, True, "none"),
|
|
("openai", "gpt-4.1", "https://proxy.example/v1"): (False, True, True, "none"),
|
|
("openai", "gpt-3.5-turbo", "https://proxy.example/v1"): (False, False, False, "none"),
|
|
("openai", "omni-model", "https://proxy.example/v1"): (False, True, False, "none"),
|
|
("openai", "deepseek-chat", "https://api.deepseek.com"): (True, True, False, "deepseek"),
|
|
("openai", "any-model-at-all", "https://api.deepseek.com/v1"): (True, True, False, "deepseek"),
|
|
}
|
|
|
|
# ProviderSpec.requires_api_key() per provider id — frozen pre-migration.
|
|
_EXPECTED_REQUIRES_API_KEY: dict[str, bool] = {
|
|
"aihubmix": True,
|
|
"anthropic": True,
|
|
"azure": True,
|
|
"bailian_coding": True,
|
|
"byteplus": True,
|
|
"byteplus_coding_plan": True,
|
|
"byteplus_coding_plan_anthropic": True,
|
|
"custom": False,
|
|
"dashscope": True,
|
|
"deepseek": True,
|
|
"gemini": True,
|
|
"github_copilot": False,
|
|
"groq": True,
|
|
"kimi_coding_anthropic": True,
|
|
"kimi_coding_openai": True,
|
|
"litellm_proxy": True,
|
|
"lm_studio": False,
|
|
"minimax": True,
|
|
"minimax_coding_anthropic": True,
|
|
"minimax_coding_openai": True,
|
|
"minimax_cn": True,
|
|
"minimax_global": True,
|
|
"minimax_openai": True,
|
|
"mimo_anthropic": True,
|
|
"mimo_openai": True,
|
|
"mistral": True,
|
|
"moonshot": True,
|
|
"ollama": False,
|
|
"openai": True,
|
|
"openai_codex": False,
|
|
"openai_responses": True,
|
|
"openrouter": True,
|
|
"ovms": False,
|
|
"qianfan": True,
|
|
"siliconflow": True,
|
|
"tencent_token_plan": True,
|
|
"tencent_token_plan_anthropic": True,
|
|
"tencent_tokenhub": True,
|
|
"tencent_tokenhub_anthropic": True,
|
|
"tencent_tokenhub_intl": True,
|
|
"tokenrhythm": True,
|
|
"vllm": False,
|
|
"volcengine": True,
|
|
"volcengine_coding_plan": True,
|
|
"volcengine_coding_plan_anthropic": True,
|
|
"zhipu": True,
|
|
}
|
|
|
|
# resolve_context_window() for a model id unknown to every layer — frozen
|
|
# pre-migration. Providers in LOCAL_RUNTIME_PROVIDERS report the local
|
|
# runtime window; everything else reports the cloud default.
|
|
_EXPECTED_UNKNOWN_CONTEXT_WINDOW: dict[str, int] = {
|
|
"aihubmix": 200000,
|
|
"anthropic": 200000,
|
|
"azure": 200000,
|
|
"bailian_coding": 200000,
|
|
"byteplus": 200000,
|
|
"byteplus_coding_plan": 200000,
|
|
"byteplus_coding_plan_anthropic": 200000,
|
|
"custom": 8192,
|
|
"dashscope": 200000,
|
|
"deepseek": 200000,
|
|
"gemini": 200000,
|
|
"github_copilot": 200000,
|
|
"groq": 200000,
|
|
"kimi_coding_anthropic": 200000,
|
|
"kimi_coding_openai": 200000,
|
|
"litellm_proxy": 200000,
|
|
"lm_studio": 8192,
|
|
"minimax": 200000,
|
|
"minimax_coding_anthropic": 200000,
|
|
"minimax_coding_openai": 200000,
|
|
"minimax_cn": 200000,
|
|
"minimax_global": 200000,
|
|
"minimax_openai": 200000,
|
|
"mimo_anthropic": 200000,
|
|
"mimo_openai": 200000,
|
|
"mistral": 200000,
|
|
"moonshot": 200000,
|
|
"ollama": 8192,
|
|
"openai": 200000,
|
|
"openai_codex": 200000,
|
|
"openai_responses": 200000,
|
|
"openrouter": 200000,
|
|
"ovms": 8192,
|
|
"qianfan": 200000,
|
|
"siliconflow": 200000,
|
|
"tencent_token_plan": 200000,
|
|
"tencent_token_plan_anthropic": 200000,
|
|
"tencent_tokenhub": 200000,
|
|
"tencent_tokenhub_anthropic": 200000,
|
|
"tencent_tokenhub_intl": 200000,
|
|
"tokenrhythm": 200000,
|
|
"vllm": 8192,
|
|
"volcengine": 200000,
|
|
"volcengine_coding_plan": 200000,
|
|
"volcengine_coding_plan_anthropic": 200000,
|
|
"zhipu": 200000,
|
|
"local": 8192,
|
|
"": 200000,
|
|
}
|
|
|
|
|
|
def test_get_capabilities_parity_full_sweep() -> None:
|
|
catalog = ModelCatalog()
|
|
for (provider, model, base_url), expected in _EXPECTED_CAPS.items():
|
|
caps = catalog.get_capabilities(model, provider_name=provider, base_url=base_url)
|
|
observed = (
|
|
caps.supports_reasoning,
|
|
caps.supports_tools,
|
|
caps.supports_vision,
|
|
caps.reasoning_format,
|
|
)
|
|
assert observed == expected, (provider, model, base_url)
|
|
assert caps.supports_streaming is True, (provider, model, base_url)
|
|
|
|
|
|
def test_sweep_covers_every_registered_provider() -> None:
|
|
swept = {provider for provider, _model, _base in _EXPECTED_CAPS}
|
|
assert set(list_provider_names()) <= swept
|
|
|
|
|
|
def test_requires_api_key_parity_every_provider() -> None:
|
|
for spec in list_provider_specs():
|
|
assert spec.requires_api_key() == _EXPECTED_REQUIRES_API_KEY[spec.provider_id], (
|
|
spec.provider_id
|
|
)
|
|
|
|
|
|
def test_unknown_model_context_window_parity_every_provider() -> None:
|
|
catalog = ModelCatalog()
|
|
for provider, window in _EXPECTED_UNKNOWN_CONTEXT_WINDOW.items():
|
|
assert catalog.resolve_context_window("model-unknown-everywhere-zz9", provider) == window, (
|
|
provider
|
|
)
|
|
|
|
|
|
def test_named_provider_sets_stay_distinct() -> None:
|
|
# Unifying membership would flip requires_api_key("vllm") — these MUST
|
|
# stay two distinct sets with an explicit superset derivation.
|
|
assert KEYLESS_PROVIDERS == frozenset({"ollama", "lm_studio", "ovms", "custom"})
|
|
assert LOCAL_RUNTIME_PROVIDERS == KEYLESS_PROVIDERS | {"vllm", "local"}
|
|
# vllm/local are local runtimes for the context heuristic but NOT keyless.
|
|
assert "vllm" in LOCAL_RUNTIME_PROVIDERS
|
|
assert "vllm" not in KEYLESS_PROVIDERS
|
|
assert get_provider_spec("vllm").requires_api_key() is False # no env_key, not keyless
|
|
# A vLLM-style deployment CAN carry auth; keyless status is not implied by
|
|
# being a local runtime. The custom endpoint is keyless AND local.
|
|
assert "custom" in KEYLESS_PROVIDERS and "custom" in LOCAL_RUNTIME_PROVIDERS
|
|
|
|
|
|
def test_local_runtime_window_matches_membership() -> None:
|
|
from opensquilla.provider.model_catalog import _LOCAL_CONTEXT_WINDOW
|
|
|
|
catalog = ModelCatalog()
|
|
for provider in LOCAL_RUNTIME_PROVIDERS:
|
|
assert (
|
|
catalog.resolve_context_window("unknown-local-model-zz9", provider)
|
|
== _LOCAL_CONTEXT_WINDOW
|
|
), provider
|
|
# A cloud provider id falls to the cloud default instead.
|
|
assert catalog.resolve_context_window("unknown-cloud-model-zz9", "openai") == (
|
|
DEFAULT_CONTEXT_WINDOW
|
|
)
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# Flag gate (decision OQ#5): anthropic/ollama early-return-empty stays the
|
|
# default for one release; flipping the flag routes both through the
|
|
# layered catalog.
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
def test_anthropic_ollama_flag_defaults_off_and_returns_empty() -> None:
|
|
from opensquilla.provider import model_catalog as model_catalog_module
|
|
from opensquilla.provider.types import ModelCapabilities
|
|
|
|
assert model_catalog_module.CATALOG_CAPABILITIES_FOR_ANTHROPIC_OLLAMA is False
|
|
catalog = ModelCatalog()
|
|
for provider, model in (("anthropic", "claude-opus-4-6"), ("ollama", "llama3.2:3b")):
|
|
assert catalog.get_capabilities(model, provider_name=provider) == ModelCapabilities()
|
|
|
|
|
|
def test_anthropic_ollama_flag_on_resolves_through_catalog(monkeypatch) -> None:
|
|
from opensquilla.provider import model_catalog as model_catalog_module
|
|
|
|
monkeypatch.setattr(
|
|
model_catalog_module, "CATALOG_CAPABILITIES_FOR_ANTHROPIC_OLLAMA", True
|
|
)
|
|
catalog = ModelCatalog()
|
|
# Snapshot knows claude-opus-4-6 (tools+vision); reasoning stays off
|
|
# because no layer names an anthropic streaming dialect yet. THIS is the
|
|
# engine-visible change the one-release gate defers: supports_vision
|
|
# flips False -> True, which stops engine-level image stripping.
|
|
claude = catalog.get_capabilities("claude-opus-4-6", provider_name="anthropic")
|
|
assert (
|
|
claude.supports_reasoning,
|
|
claude.supports_tools,
|
|
claude.supports_vision,
|
|
claude.reasoning_format,
|
|
) == (False, True, True, "none")
|
|
# An unqualified local model is unknown to every layer -> synthesized
|
|
# floor, which happens to match the historical empty capabilities.
|
|
local = catalog.get_capabilities("llama3.2:3b", provider_name="ollama")
|
|
assert (
|
|
local.supports_reasoning,
|
|
local.supports_tools,
|
|
local.supports_vision,
|
|
local.reasoning_format,
|
|
) == (False, True, False, "none")
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# Ladder ordering that resolve_entry cannot express is kept as code; freeze
|
|
# it here: a live OpenRouter reasoning hit outranks the api.openai.com host
|
|
# guard, and host trust never leaks to data rows.
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
def _catalog_with_live_row(model_id: str, *, reasoning: bool) -> ModelCatalog:
|
|
catalog = ModelCatalog()
|
|
catalog._populate_from_data(
|
|
[
|
|
{
|
|
"id": model_id,
|
|
"name": "Synthetic live row",
|
|
"context_length": 128_000,
|
|
"top_provider": {"max_completion_tokens": 32_768},
|
|
"supported_parameters": ["tools", "reasoning"] if reasoning else ["tools"],
|
|
"architecture": {"input_modalities": ["text", "image"]},
|
|
}
|
|
]
|
|
)
|
|
return catalog
|
|
|
|
|
|
def test_live_reasoning_hit_outranks_openai_host_guard() -> None:
|
|
catalog = _catalog_with_live_row("gpt-5-proxy-row", reasoning=True)
|
|
caps = catalog.get_capabilities(
|
|
"gpt-5-proxy-row",
|
|
provider_name="openai",
|
|
base_url="https://api.openai.com/v1",
|
|
)
|
|
# Pre-migration order: the live-catalog reasoning branch ran BEFORE the
|
|
# api.openai.com prefix branch, so the OpenRouter dialect wins.
|
|
assert caps.supports_reasoning is True
|
|
assert caps.reasoning_format == "openrouter"
|
|
assert caps.supports_vision is True # live row carries image modality
|
|
|
|
|
|
def test_live_non_reasoning_hit_falls_through_to_layers() -> None:
|
|
catalog = _catalog_with_live_row("vendor/plain-model", reasoning=False)
|
|
caps = catalog.get_capabilities("vendor/plain-model", provider_name="openrouter")
|
|
# Same as the legacy final fallback: live tools/vision, no reasoning.
|
|
assert caps.supports_reasoning is False
|
|
assert caps.reasoning_format == "none"
|
|
assert caps.supports_tools is True
|
|
assert caps.supports_vision is True
|
|
|
|
|
|
def test_deepseek_base_url_sniff_stays_code_for_openai_provider_only() -> None:
|
|
catalog = ModelCatalog()
|
|
sniffed = catalog.get_capabilities(
|
|
"some-model", provider_name="openai", base_url="https://api.deepseek.com/v1"
|
|
)
|
|
assert sniffed.supports_reasoning is True
|
|
assert sniffed.reasoning_format == "deepseek"
|
|
# Another provider id with the same base URL does not trigger the sniff.
|
|
other = catalog.get_capabilities(
|
|
"some-model", provider_name="groq", base_url="https://api.deepseek.com/v1"
|
|
)
|
|
assert other.supports_reasoning is False
|
|
assert other.reasoning_format == "none"
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# New, deliberate capability of the migration: the user-override layer now
|
|
# reaches get_capabilities (highest authority), so operators can fix a
|
|
# wrong capability row without waiting for a release.
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
def test_user_overrides_now_reach_get_capabilities() -> None:
|
|
catalog = ModelCatalog()
|
|
catalog.set_user_overrides(
|
|
{
|
|
"dashscope/qwen3-max": {"supports_reasoning": False},
|
|
"brand-new-model": {
|
|
"supports_reasoning": True,
|
|
"reasoning_format": "deepseek",
|
|
"supports_vision": True,
|
|
},
|
|
}
|
|
)
|
|
# An explicit False beats the transcribed ladder row...
|
|
downgraded = catalog.get_capabilities("qwen3-max", provider_name="dashscope")
|
|
assert downgraded.supports_reasoning is False
|
|
assert downgraded.reasoning_format == "none"
|
|
# ...and a full user row can enable reasoning for an unknown model,
|
|
# because it names the dialect itself.
|
|
upgraded = catalog.get_capabilities("brand-new-model", provider_name="siliconflow")
|
|
assert upgraded.supports_reasoning is True
|
|
assert upgraded.reasoning_format == "deepseek"
|
|
assert upgraded.supports_vision is True
|
|
|
|
|
|
def test_reasoning_without_dialect_is_never_invented() -> None:
|
|
# A user row (or snapshot row) claiming supports_reasoning WITHOUT a
|
|
# reasoning_format must not enable reasoning: the adapter refuses to
|
|
# invent a streaming dialect (the legacy fallback's exact semantics).
|
|
catalog = ModelCatalog()
|
|
catalog.set_user_overrides({"formatless-model": {"supports_reasoning": True}})
|
|
caps = catalog.get_capabilities("formatless-model", provider_name="siliconflow")
|
|
assert caps.supports_reasoning is False
|
|
assert caps.reasoning_format == "none"
|