chore: import upstream snapshot with attribution

providers/inference/logo.svg

@@ -0,0 +1,3 @@
+<svg width="24" height="24" viewBox="0 0 40 40" xmlns="http://www.w3.org/2000/svg">
+<path d="M34.625 33.7119H27.4734V6.28124H34.625V33.7119ZM23.703 33.7119H18.6296V6.28124H23.703V33.7119ZM15.2674 33.7119H11.3771V6.28124H15.2674V33.7119ZM8.15814 33.7119H5.375V6.28124H8.15814V33.7119Z" fill="currentColor"/>
+</svg>

providers/inference/models/google/gemma-3.toml

@@ -0,0 +1,23 @@
+name = "Google Gemma 3"
+description = "Open Gemma instruction model for efficient chat and self-hosted deployments"
+family = "gemma"
+release_date = "2025-01-01"
+last_updated = "2025-01-01"
+attachment = true
+reasoning = false
+temperature = true
+knowledge = "2024-12"
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.15
+output = 0.30
+
+[limit]
+context = 125_000
+output = 4_096
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]

providers/inference/models/meta/llama-3.1-8b-instruct.toml

@@ -0,0 +1,23 @@
+name = "Llama 3.1 8B Instruct"
+description = "Open Llama instruction model for multilingual chat, reasoning, and coding"
+family = "llama"
+release_date = "2025-01-01"
+last_updated = "2025-01-01"
+attachment = false
+reasoning = false
+temperature = true
+knowledge = "2023-12"
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.025
+output = 0.025
+
+[limit]
+context = 16_000
+output = 4_096
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/inference/models/meta/llama-3.2-11b-vision-instruct.toml

@@ -0,0 +1,23 @@
+name = "Llama 3.2 11B Vision Instruct"
+description = "Open Llama multimodal model for image understanding and text reasoning"
+family = "llama"
+release_date = "2025-01-01"
+last_updated = "2025-01-01"
+attachment = true
+reasoning = false
+temperature = true
+knowledge = "2023-12"
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.055
+output = 0.055
+
+[limit]
+context = 16_000
+output = 4_096
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]

providers/inference/models/meta/llama-3.2-1b-instruct.toml

@@ -0,0 +1,23 @@
+name = "Llama 3.2 1B Instruct"
+description = "Open Llama instruction model for multilingual chat, reasoning, and coding"
+family = "llama"
+release_date = "2025-01-01"
+last_updated = "2025-01-01"
+attachment = false
+reasoning = false
+temperature = true
+knowledge = "2023-12"
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.01
+output = 0.01
+
+[limit]
+context = 16_000
+output = 4_096
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/inference/models/meta/llama-3.2-3b-instruct.toml

@@ -0,0 +1,23 @@
+name = "Llama 3.2 3B Instruct"
+description = "Open Llama instruction model for multilingual chat, reasoning, and coding"
+family = "llama"
+release_date = "2025-01-01"
+last_updated = "2025-01-01"
+attachment = false
+reasoning = false
+temperature = true
+knowledge = "2023-12"
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.02
+output = 0.02
+
+[limit]
+context = 16_000
+output = 4_096
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/inference/models/mistral/mistral-nemo-12b-instruct.toml

@@ -0,0 +1,23 @@
+name = "Mistral Nemo 12B Instruct"
+description = "Mistral model for multilingual chat, reasoning, and tool-assisted workflows"
+family = "mistral-nemo"
+release_date = "2025-01-01"
+last_updated = "2025-01-01"
+attachment = false
+reasoning = false
+temperature = true
+knowledge = "2024-12"
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.038
+output = 0.10
+
+[limit]
+context = 16_000
+output = 4_096
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/inference/models/osmosis/osmosis-structure-0.6b.toml

@@ -0,0 +1,23 @@
+name = "Osmosis Structure 0.6B"
+description = "Open-weight instruction model for adaptable chat and self-hosted production workloads"
+family = "osmosis"
+release_date = "2025-01-01"
+last_updated = "2025-01-01"
+attachment = false
+reasoning = false
+temperature = true
+knowledge = "2024-12"
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.10
+output = 0.50
+
+[limit]
+context = 4_000
+output = 2_048
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/inference/models/qwen/qwen-2.5-7b-vision-instruct.toml

@@ -0,0 +1,23 @@
+name = "Qwen 2.5 7B Vision Instruct"
+description = "Qwen vision-language model for visual reasoning, documents, and agent tasks"
+family = "qwen"
+release_date = "2025-01-01"
+last_updated = "2025-01-01"
+attachment = true
+reasoning = false
+temperature = true
+knowledge = "2024-12"
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.20
+output = 0.20
+
+[limit]
+context = 125_000
+output = 4_096
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]

providers/inference/models/qwen/qwen3-embedding-4b.toml

@@ -0,0 +1,23 @@
+name = "Qwen 3 Embedding 4B"
+description = "Embedding model for semantic search, retrieval, clustering, and ranking pipelines"
+family = "qwen"
+release_date = "2025-01-01"
+last_updated = "2025-01-01"
+attachment = false
+reasoning = false
+temperature = false
+knowledge = "2024-12"
+tool_call = false
+open_weights = true
+
+[cost]
+input = 0.01
+output = 0.00
+
+[limit]
+context = 32_000
+output = 2_048
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/inference/provider.toml

@@ -0,0 +1,9 @@
+name = "Inference"
+env = ["INFERENCE_API_KEY"]
+npm = "@ai-sdk/openai-compatible"
+# Reasoning HTTP format (accessed 2026-06-25): POST /v1/chat/completions. The
+# provider's supported-parameter table documents no toggle, effort, or numeric
+# reasoning-token budget field; it directs users to request unlisted fields.
+# https://docs.inference.net/api/api-quickstart
+api = "https://inference.net/v1"
+doc = "https://inference.net/models"