chore: import upstream snapshot with attribution

2026-07-13 12:28:55 +08:00
commit db42b91b75
6397 changed files with 146012 additions and 0 deletions
@@ -0,0 +1,3 @@
+<svg width="24" height="24" viewBox="0 0 24 24" fill="none" xmlns="http://www.w3.org/2000/svg">
+  <path d="M8.69514 3.98663C3.44358 6.10184 1.69306 13.7604 5.55879 17.6261L7.60106 19.5954L9.86215 17.4802C11.1021 16.2403 12.415 14.5627 12.7797 13.6145C13.2173 12.5204 13.9467 12.0098 14.6761 12.1557C16.4266 12.5204 16.7913 14.8544 15.1137 15.4379C13.1444 16.0944 10.5186 18.6472 10.5186 19.8872C10.5186 21.7106 15.6243 20.3977 18.1771 17.8449C20.0006 15.9485 20.3653 15.0732 20.3653 12.0098C20.3653 9.09231 20.0006 7.99823 18.4689 6.53947L16.5725 4.64307L14.2384 6.61241C12.9256 7.70648 11.5397 9.38406 11.175 10.3323C10.7374 11.4993 10.081 12.0098 9.27864 11.864C7.52813 11.4993 7.16343 9.16524 8.84101 8.65468C10.7374 7.92529 14.1655 4.64307 13.7279 3.84075C13.2902 3.11137 10.6645 3.18431 8.69514 3.98663Z" fill="currentColor"/>
+</svg>
@@ -0,0 +1,22 @@
+name = "Qwen3 235B A22B Instruct 2507"
+description = "Qwen instruction model for multilingual chat, reasoning, and tool use"
+family = "qwen"
+release_date = "2025-08-23"
+last_updated = "2025-08-23"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.20
+output = 0.30
+
+[limit]
+context = 262_144
+output = 131_072
+
+[modalities]
+input = ["text"]
+output = ["text"]
@@ -0,0 +1,23 @@
+name = "Qwen3 235B A22B Thinking 2507"
+description = "Qwen reasoning model for deliberate problem solving, math, and coding"
+family = "qwen"
+release_date = "2025-08-23"
+last_updated = "2025-08-23"
+attachment = false
+reasoning = true
+reasoning_options = []
+temperature = true
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.20
+output = 0.60
+
+[limit]
+context = 262_144
+output = 131_072
+
+[modalities]
+input = ["text"]
+output = ["text"]
@@ -0,0 +1,22 @@
+name = "Qwen3 Coder 480B A35B Instruct"
+description = "Qwen coding model for software agents, repository edits, and code reasoning"
+family = "qwen"
+release_date = "2025-08-23"
+last_updated = "2025-08-23"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = true
+open_weights = false
+
+[cost]
+input = 0.20
+output = 0.80
+
+[limit]
+context = 262_144
+output = 262_144
+
+[modalities]
+input = ["text"]
+output = ["text"]
@@ -0,0 +1,23 @@
+name = "DeepSeek R1 0528"
+description = "DeepSeek reasoning model for multi-step analysis, math, coding, and tools"
+family = "deepseek-thinking"
+release_date = "2025-08-23"
+last_updated = "2025-08-23"
+attachment = false
+reasoning = true
+reasoning_options = []
+temperature = true
+tool_call = true
+open_weights = false
+
+[cost]
+input = 0.5
+output = 2.15
+
+[limit]
+context = 75_000
+output = 163_840
+
+[modalities]
+input = ["text"]
+output = ["text"]
@@ -0,0 +1,22 @@
+name = "DeepSeek V3 0324"
+description = "DeepSeek chat model for instruction following, coding, and analysis"
+family = "deepseek"
+release_date = "2025-08-23"
+last_updated = "2025-08-23"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = true
+open_weights = false
+
+[cost]
+input = 0.20
+output = 0.80
+
+[limit]
+context = 75_000
+output = 163_840
+
+[modalities]
+input = ["text"]
+output = ["text"]
@@ -0,0 +1,23 @@
+name = "DeepSeek V3.1"
+description = "DeepSeek chat model for instruction following, coding, and analysis"
+family = "deepseek"
+release_date = "2025-08-23"
+last_updated = "2025-08-23"
+attachment = false
+reasoning = true
+reasoning_options = []
+temperature = true
+tool_call = true
+open_weights = false
+
+[cost]
+input = 0.20
+output = 0.80
+
+[limit]
+context = 75_000
+output = 163_840
+
+[modalities]
+input = ["text"]
+output = ["text"]
@@ -0,0 +1,23 @@
+name = "GPT OSS 120B"
+description = "Open-weight GPT model for self-hosted reasoning and instruction-following workloads"
+family = "gpt-oss"
+release_date = "2025-08-23"
+last_updated = "2025-08-23"
+attachment = false
+reasoning = true
+reasoning_options = []
+temperature = true
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.10
+output = 0.50
+
+[limit]
+context = 131_072
+output = 32_768
+
+[modalities]
+input = ["text"]
+output = ["text"]
@@ -0,0 +1,22 @@
+name = "GLM 4.5 Air"
+description = "Efficient GLM model for fast reasoning, coding, and agent workflows"
+family = "glm-air"
+release_date = "2025-07-28"
+last_updated = "2025-07-28"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.10
+output = 0.50
+
+[limit]
+context = 131_072
+output = 131_072
+
+[modalities]
+input = ["text"]
+output = ["text"]
@@ -0,0 +1,23 @@
+name = "GLM 4.5 FP8"
+description = "Flagship GLM model for hybrid reasoning, coding, and agentic engineering"
+family = "glm"
+release_date = "2025-07-28"
+last_updated = "2025-07-28"
+attachment = false
+reasoning = true
+reasoning_options = []
+temperature = true
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.20
+output = 0.80
+
+[limit]
+context = 131_072
+output = 131_072
+
+[modalities]
+input = ["text"]
+output = ["text"]
@@ -0,0 +1,12 @@
+name = "submodel"
+env = ["SUBMODEL_INSTAGEN_ACCESS_KEY"]
+npm = "@ai-sdk/openai-compatible"
+# Reasoning HTTP format (accessed 2026-06-25):
+# InstaGen documents POST /v1/chat/completions but no caller reasoning toggle,
+# effort field, token budget, values, bounds, sentinel, or reasoning response
+# path. Its model catalog labels models as reasoning/general only.
+# Sources:
+# https://submodel.gitbook.io/docs/instagen/api-reference
+# https://submodel.gitbook.io/docs/instagen/overview-1/available-models
+api = "https://llm.submodel.ai/v1"
+doc = "https://submodel.gitbook.io"