chore: import upstream snapshot with attribution

providers/nvidia/models/abacusai/dracarys-llama-3_1-70b-instruct.toml

@@ -0,0 +1,21 @@
+name = "dracarys-llama-3.1-70b-instruct"
+description = "Open Llama instruction model for multilingual chat, reasoning, and coding"
+release_date = "2024-09-11"
+last_updated = "2025-05-22"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 128_000
+output = 8_192
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/baai/bge-m3.toml

@@ -0,0 +1,22 @@
+name = "BGE M3"
+description = "Flagship model for demanding analysis, coding, and production agent workflows"
+family = "bge"
+release_date = "2024-01-30"
+last_updated = "2026-04-30"
+attachment = false
+reasoning = false
+temperature = false
+tool_call = false
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 8_192
+output = 1_024
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/black-forest-labs/flux.1-dev.toml

@@ -0,0 +1,23 @@
+name = "FLUX.1-dev"
+description = "Image model for prompt-driven generation, editing, and visual design workflows"
+family = "flux"
+release_date = "2024-08-01"
+last_updated = "2025-09-05"
+attachment = false
+reasoning = false
+temperature = true
+knowledge = "2024-08"
+tool_call = false
+open_weights = false
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 4096
+output = 0
+
+[modalities]
+input = ["text"]
+output = ["image"]

providers/nvidia/models/black-forest-labs/flux_1-kontext-dev.toml

@@ -0,0 +1,21 @@
+name = "FLUX.1-Kontext-dev"
+description = "Image model for prompt-driven generation, editing, and visual design workflows"
+release_date = "2025-08-12"
+last_updated = "2025-08-12"
+attachment = true
+reasoning = false
+temperature = false
+tool_call = false
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 40_960
+output = 40_960
+
+[modalities]
+input = ["text", "image"]
+output = ["image"]

providers/nvidia/models/black-forest-labs/flux_1-schnell.toml

@@ -0,0 +1,24 @@
+name = "FLUX.1-schnell"
+description = "Image model for prompt-driven generation, editing, and visual design workflows"
+release_date = "2024-08-01"
+last_updated = "2026-02-04"
+attachment = false
+reasoning = false
+temperature = false
+knowledge = "2024-07"
+tool_call = false
+structured_output = false
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 77
+input = 77
+output = 0
+
+[modalities]
+input = ["text"]
+output = ["image"]

providers/nvidia/models/black-forest-labs/flux_2-klein-4b.toml

@@ -0,0 +1,23 @@
+name = "FLUX.2 Klein 4B"
+description = "Image model for prompt-driven generation, editing, and visual design workflows"
+family = "flux"
+release_date = "2026-01-14"
+last_updated = "2026-01-31"
+attachment = false
+reasoning = false
+temperature = true
+knowledge = "2025-06"
+tool_call = false
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 40_960
+output = 40_960
+
+[modalities]
+input = ["image", "text"]
+output = ["image"]

providers/nvidia/models/bytedance/seed-oss-36b-instruct.toml

@@ -0,0 +1,23 @@
+name = "ByteDance-Seed/Seed-OSS-36B-Instruct"
+description = "Tool-capable chat model for instruction following and agentic application workflows"
+family = "seed"
+release_date = "2025-09-04"
+last_updated = "2025-11-25"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = true
+structured_output = true
+open_weights = false
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 262_000
+output = 262_000
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/deepseek-ai/deepseek-v4-flash.toml

@@ -0,0 +1,17 @@
+base_model = "deepseek/deepseek-v4-flash"
+
+# NIM Chat schema: `reasoning_effort = none|high|max`; `none` disables thinking.
+# https://docs.api.nvidia.com/nim/reference/deepseek-ai-deepseek-v4-flash-infer
+reasoning_options = [{ type = "effort", values = ["none", "high", "max"] }]
+
+[interleaved]
+field = "reasoning_content"
+
+[cost]
+input = 0.14
+output = 0.28
+cache_read = 0.0028
+
+[limit]
+context = 1_048_576
+output = 393_216

providers/nvidia/models/deepseek-ai/deepseek-v4-pro.toml

@@ -0,0 +1,17 @@
+base_model = "deepseek/deepseek-v4-pro"
+
+# NIM Chat schema: `reasoning_effort = none|high|max`; `none` disables thinking.
+# https://docs.api.nvidia.com/nim/reference/deepseek-ai-deepseek-v4-pro-infer
+reasoning_options = [{ type = "effort", values = ["none", "high", "max"] }]
+
+[interleaved]
+field = "reasoning_content"
+
+[cost]
+input = 0.435
+output = 0.87
+cache_read = 0.003625
+
+[limit]
+context = 1_048_576
+output = 393_216

providers/nvidia/models/google/gemma-2-2b-it.toml

@@ -0,0 +1,22 @@
+name = "Gemma 2 2b It"
+description = "Open Gemma instruction model for efficient chat and self-hosted deployments"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = true
+structured_output = true
+release_date = "2024-07-16"
+last_updated = "2024-07-16"
+open_weights = true
+
+[cost]
+input = 0.00
+output = 0.00
+
+[limit]
+context = 128000
+output = 4096
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/google/gemma-3n-e2b-it.toml

@@ -0,0 +1,23 @@
+name = "Gemma 3n E2b It"
+description = "Open Gemma instruction model for efficient chat and self-hosted deployments"
+attachment = true
+reasoning = false
+temperature = true
+knowledge = "2024-06"
+tool_call = true
+structured_output = true
+release_date = "2025-06-12"
+last_updated = "2025-06-12"
+open_weights = true
+
+[cost]
+input = 0.00
+output = 0.00
+
+[limit]
+context = 128000
+output = 4096
+
+[modalities]
+input = ["text","image"]
+output = ["text"]

providers/nvidia/models/google/gemma-3n-e4b-it.toml

@@ -0,0 +1,23 @@
+name = "Gemma 3n E4b It"
+description = "Open Gemma instruction model for efficient chat and self-hosted deployments"
+attachment = true
+reasoning = false
+temperature = true
+knowledge = "2024-06"
+tool_call = true
+structured_output = true
+release_date = "2025-06-03"
+last_updated = "2025-06-03"
+open_weights = true
+
+[cost]
+input = 0.00
+output = 0.00
+
+[limit]
+context = 128000
+output = 4096
+
+[modalities]
+input = ["text","image"]
+output = ["text"]

providers/nvidia/models/google/gemma-4-31b-it.toml

@@ -0,0 +1,24 @@
+name = "Gemma-4-31B-IT"
+description = "Open Gemma instruction model for efficient chat and self-hosted deployments"
+family = "gemma"
+release_date = "2026-04-02"
+last_updated = "2026-04-02"
+attachment = true
+reasoning = true
+reasoning_options = [{ type = "toggle" }]
+temperature = true
+knowledge = "2025-01"
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 256000
+output = 16384
+
+[modalities]
+input = ["text", "image", "video"]
+output = ["text"]

providers/nvidia/models/google/google-paligemma.toml

@@ -0,0 +1,21 @@
+name = "paligemma"
+description = "Gemini multimodal model for text, image, audio, video, and document tasks"
+release_date = "2024-05-14"
+last_updated = "2024-08-26"
+attachment = true
+reasoning = false
+temperature = true
+tool_call = false
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 128_000
+output = 8_192
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]

providers/nvidia/models/meta/esm2-650m.toml

@@ -0,0 +1,21 @@
+name = "esm2-650m"
+description = "Open Llama instruction model for multilingual chat, reasoning, and coding"
+release_date = "2024-08-29"
+last_updated = "2025-03-10"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = false
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 128_000
+output = 8_192
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/meta/esmfold.toml

@@ -0,0 +1,21 @@
+name = "esmfold"
+description = "Open Llama instruction model for multilingual chat, reasoning, and coding"
+release_date = "2024-03-15"
+last_updated = "2025-06-12"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = false
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 128_000
+output = 8_192
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/meta/llama-3.1-70b-instruct.toml

@@ -0,0 +1,22 @@
+name = "Llama 3.1 70b Instruct"
+description = "Open Llama instruction model for multilingual chat, reasoning, and coding"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = true
+structured_output = true
+release_date = "2024-07-16"
+last_updated = "2024-07-16"
+open_weights = true
+
+[cost]
+input = 0.00
+output = 0.00
+
+[limit]
+context = 128000
+output = 4096
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/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 = 0.0
+output = 0.0
+
+[limit]
+context = 16_000
+output = 4_096
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/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"
+attachment = true
+reasoning = false
+temperature = true
+knowledge = "2023-12"
+tool_call = true
+structured_output = true
+release_date = "2024-09-18"
+last_updated = "2024-09-18"
+open_weights = true
+
+[cost]
+input = 0.00
+output = 0.00
+
+[limit]
+context = 128000
+output = 4096
+
+[modalities]
+input = ["text","image"]
+output = ["text"]

providers/nvidia/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"
+attachment = false
+reasoning = false
+temperature = true
+knowledge = "2023-12"
+tool_call = true
+structured_output = true
+release_date = "2024-09-18"
+last_updated = "2024-09-18"
+open_weights = true
+
+[cost]
+input = 0.00
+output = 0.00
+
+[limit]
+context = 128000
+output = 4096
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/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 = "2024-09-18"
+last_updated = "2024-09-18"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = false
+structured_output = true
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 32_768
+output = 32_000
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/meta/llama-3.2-90b-vision-instruct.toml

@@ -0,0 +1,23 @@
+name = "Llama-3.2-90B-Vision-Instruct"
+description = "Open Llama multimodal model for image understanding and text reasoning"
+family = "llama"
+release_date = "2024-09-25"
+last_updated = "2024-09-25"
+attachment = true
+reasoning = false
+temperature = true
+knowledge = "2023-12"
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 128_000
+output = 8_192
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]

providers/nvidia/models/meta/llama-3.3-70b-instruct.toml

@@ -0,0 +1,22 @@
+name = "Llama 3.3 70b Instruct"
+description = "Open Llama instruction model for multilingual chat, reasoning, and coding"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = true
+structured_output = true
+release_date = "2024-11-26"
+last_updated = "2024-11-26"
+open_weights = true
+
+[cost]
+input = 0.00
+output = 0.00
+
+[limit]
+context = 128000
+output = 4096
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/meta/llama-4-maverick-17b-128e-instruct.toml

@@ -0,0 +1,23 @@
+name = "Llama 4 Maverick 17b 128e Instruct"
+description = "Open multimodal Llama model for strong reasoning and fast responses"
+attachment = true
+reasoning = false
+temperature = true
+knowledge = "2024-02"
+tool_call = true
+structured_output = true
+release_date = "2025-04-01"
+last_updated = "2025-04-01"
+open_weights = true
+
+[cost]
+input = 0.00
+output = 0.00
+
+[limit]
+context = 128000
+output = 4096
+
+[modalities]
+input = ["text","image"]
+output = ["text"]

providers/nvidia/models/meta/llama-guard-4-12b.toml

@@ -0,0 +1,22 @@
+name = "Llama Guard 4 12B"
+description = "Safety model for policy screening, moderation, and risk-aware routing workflows"
+family = "llama"
+release_date = "2025-04-05"
+last_updated = "2026-04-30"
+attachment = true
+reasoning = false
+temperature = true
+tool_call = false
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 128_000
+output = 16_384
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]

providers/nvidia/models/microsoft/phi-4-mini-instruct.toml

@@ -0,0 +1,23 @@
+name = "Phi-4-Mini"
+description = "Efficient model for low-latency assistance, extraction, and routine automation"
+family = "phi"
+release_date = "2024-12-01"
+last_updated = "2025-09-05"
+attachment = false
+reasoning = false
+temperature = true
+knowledge = "2024-12"
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 131_072
+output = 8_192
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/microsoft/phi-4-multimodal-instruct.toml

@@ -0,0 +1,22 @@
+name = "Phi 4 Multimodal"
+description = "General-purpose chat model for instruction following, writing, and analysis"
+release_date = "2025-07-26"
+last_updated = "2025-07-26"
+attachment = false
+reasoning = false
+tool_call = false
+structured_output = false
+open_weights = false
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 128_000
+input = 128_000
+output = 16_384
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/minimaxai/minimax-m2.7.toml

@@ -0,0 +1,23 @@
+name = "MiniMax-M2.7"
+description = "MiniMax model for chat, coding, office work, and agentic tasks"
+family = "minimax"
+release_date = "2026-03-18"
+last_updated = "2026-04-11"
+attachment = false
+reasoning = true
+reasoning_options = []
+temperature = true
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 204800
+output = 131072
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/minimaxai/minimax-m3.toml

@@ -0,0 +1,10 @@
+base_model = "minimax/MiniMax-M3"
+reasoning_options = [{ type = "toggle" }]
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 1_000_000
+output = 16_384

providers/nvidia/models/mistralai/magistral-small-2506.toml

@@ -0,0 +1,22 @@
+name = "Magistral Small 2506"
+description = "Mistral reasoning model for transparent analysis, math, and complex decisions"
+release_date = "2025-09-25"
+last_updated = "2025-09-25"
+attachment = false
+reasoning = false
+tool_call = false
+structured_output = false
+open_weights = false
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 32_768
+input = 32_768
+output = 32_768
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/mistralai/mistral-7b-instruct-v03.toml

@@ -0,0 +1,22 @@
+name = "Mistral-7B-Instruct-v0.3"
+description = "Mistral model for multilingual chat, reasoning, and tool-assisted workflows"
+release_date = "2025-04-01"
+last_updated = "2025-04-01"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = true
+structured_output = true
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 65_536
+output = 65_536
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/mistralai/mistral-large-3-675b-instruct-2512.toml

@@ -0,0 +1,24 @@
+name = "Mistral Large 3 675B Instruct 2512"
+description = "Flagship Mistral model for advanced reasoning, coding, and multilingual work"
+family = "mistral-large"
+attachment = true
+reasoning = false
+tool_call = true
+structured_output = true
+temperature = true
+knowledge = "2025-01"
+release_date = "2025-12-02"
+last_updated = "2025-12-02"
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 262_144
+output = 262_144
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]

providers/nvidia/models/mistralai/mistral-medium-3-instruct.toml

@@ -0,0 +1,23 @@
+name = "Mistral Medium 3"
+description = "Mistral model for multilingual chat, reasoning, and tool-assisted workflows"
+family = "mistral-medium"
+release_date = "2025-09-25"
+last_updated = "2025-09-25"
+attachment = true
+reasoning = false
+tool_call = false
+structured_output = false
+open_weights = false
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 131_072
+input = 131_072
+output = 32_768
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]

providers/nvidia/models/mistralai/mistral-nemotron.toml

@@ -0,0 +1,21 @@
+name = "mistral-nemotron"
+base_model = "nvidia/mistral-nemotron"
+release_date = "2025-06-11"
+last_updated = "2025-06-12"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 128_000
+output = 8_192
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/mistralai/mistral-small-4-119b-2603.toml

@@ -0,0 +1,23 @@
+name = "mistral-small-4-119b-2603"
+description = "Efficient Mistral model for fast chat, extraction, and production assistants"
+release_date = "2026-03-16"
+last_updated = "2026-03-16"
+attachment = true
+reasoning = true
+reasoning_options = [{ type = "effort", values = ["none", "high"] }]
+temperature = true
+tool_call = true
+structured_output = true
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 128_000
+output = 8_192
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]

providers/nvidia/models/mistralai/mixtral-8x22b-instruct.toml

@@ -0,0 +1,21 @@
+name = "Mistral: Mixtral 8x22B Instruct"
+description = "Mistral model for multilingual chat, reasoning, and tool-assisted workflows"
+release_date = "2024-04-17"
+last_updated = "2024-04-17"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 65_536
+output = 13_108
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/mistralai/mixtral-8x7b-instruct.toml

@@ -0,0 +1,21 @@
+name = "Mistral: Mixtral 8x7B Instruct"
+description = "Mistral model for multilingual chat, reasoning, and tool-assisted workflows"
+release_date = "2023-12-10"
+last_updated = "2026-03-15"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 32_768
+output = 16_384
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/moonshotai/kimi-k2-instruct-0905.toml

@@ -0,0 +1,24 @@
+name = "Kimi K2 0905"
+status = "deprecated"
+description = "Kimi model for long-context chat, coding, and agentic reasoning"
+family = "kimi-k2"
+release_date = "2025-09-05"
+last_updated = "2025-09-05"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = true
+knowledge = "2024-10"
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 262_144
+output = 262_144
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/moonshotai/kimi-k2.6.toml

@@ -0,0 +1,29 @@
+name = "Kimi K2.6"
+status = "deprecated"
+description = "Kimi multimodal agent model for visual understanding, coding, and planning"
+family = "kimi-k2"
+release_date = "2026-04-21"
+last_updated = "2026-04-21"
+attachment = true
+reasoning = true
+reasoning_options = [{ type = "effort", values = ["none", "minimal", "low", "medium", "high", "xhigh", "max"] }]
+structured_output = true
+temperature = true
+tool_call = true
+knowledge = "2025-01"
+open_weights = true
+
+[interleaved]
+field = "reasoning_content"
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 262_144
+output = 262_144
+
+[modalities]
+input = ["text", "image", "video"]
+output = ["text"]

providers/nvidia/models/nvidia/active-speaker-detection.toml

@@ -0,0 +1,21 @@
+name = "Active Speaker Detection"
+description = "Nemotron multimodal model for visual reasoning and agentic AI workflows"
+release_date = "2026-04-16"
+last_updated = "2026-04-16"
+attachment = true
+reasoning = false
+temperature = false
+tool_call = false
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 0
+output = 4_096
+
+[modalities]
+input = ["video"]
+output = ["text"]

providers/nvidia/models/nvidia/bevformer.toml

@@ -0,0 +1,21 @@
+name = "bevformer"
+description = "Nemotron multimodal model for visual reasoning and agentic AI workflows"
+release_date = "2025-03-18"
+last_updated = "2025-07-20"
+attachment = true
+reasoning = false
+temperature = true
+tool_call = false
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 128_000
+output = 8_192
+
+[modalities]
+input = ["video"]
+output = ["text"]

providers/nvidia/models/nvidia/cosmos-predict1-5b.toml

@@ -0,0 +1,21 @@
+name = "cosmos-predict1-5b"
+description = "Video model for prompt-guided generation, editing, and motion workflows"
+release_date = "2025-03-18"
+last_updated = "2025-03-18"
+attachment = true
+reasoning = false
+temperature = false
+tool_call = false
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 0
+output = 4_096
+
+[modalities]
+input = ["text", "image", "video"]
+output = ["video"]

providers/nvidia/models/nvidia/cosmos-transfer1-7b.toml

@@ -0,0 +1,21 @@
+name = "cosmos-transfer1-7b"
+description = "Video model for prompt-guided generation, editing, and motion workflows"
+release_date = "2025-06-13"
+last_updated = "2025-06-30"
+attachment = true
+reasoning = false
+temperature = false
+tool_call = false
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 0
+output = 4_096
+
+[modalities]
+input = ["text", "image", "video"]
+output = ["video"]

providers/nvidia/models/nvidia/cosmos-transfer2_5-2b.toml

@@ -0,0 +1,21 @@
+name = "cosmos-transfer2.5-2b"
+description = "Video model for prompt-guided generation, editing, and motion workflows"
+release_date = "2026-02-26"
+last_updated = "2026-02-26"
+attachment = true
+reasoning = false
+temperature = false
+tool_call = false
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 0
+output = 4_096
+
+[modalities]
+input = ["text", "image", "video"]
+output = ["video"]

providers/nvidia/models/nvidia/gliner-pii.toml

@@ -0,0 +1,21 @@
+name = "gliner-pii"
+description = "Nemotron model for efficient reasoning, coding, and specialized AI agents"
+release_date = "2026-03-03"
+last_updated = "2026-03-03"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = false
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 128_000
+output = 4_096
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/nvidia/llama-3_1-nemotron-safety-guard-8b-v3.toml

@@ -0,0 +1,21 @@
+name = "llama-3.1-nemotron-safety-guard-8b-v3"
+base_model = "nvidia/llama-3.1-nemotron-safety-guard-8b-v3"
+release_date = "2025-10-28"
+last_updated = "2025-10-28"
+attachment = false
+reasoning = false
+temperature = false
+tool_call = false
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 128_000
+output = 4_096
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/nvidia/llama-3_2-nemoretriever-300m-embed-v1.toml

@@ -0,0 +1,21 @@
+name = "llama-3_2-nemoretriever-300m-embed-v1"
+description = "Embedding model for semantic search, retrieval, clustering, and ranking pipelines"
+release_date = "2025-07-24"
+last_updated = "2025-07-24"
+attachment = false
+reasoning = false
+temperature = false
+tool_call = false
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 32_768
+output = 2_048
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/nvidia/llama-nemotron-embed-vl-1b-v2.toml

@@ -0,0 +1,21 @@
+name = "llama-nemotron-embed-vl-1b-v2"
+base_model = "nvidia/llama-nemotron-embed-vl-1b-v2"
+release_date = "2026-02-10"
+last_updated = "2026-02-10"
+attachment = true
+reasoning = false
+temperature = false
+tool_call = false
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 32_768
+output = 2_048
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]

providers/nvidia/models/nvidia/llama-nemotron-rerank-vl-1b-v2.toml

@@ -0,0 +1,21 @@
+name = "llama-nemotron-rerank-vl-1b-v2"
+base_model = "nvidia/llama-nemotron-rerank-vl-1b-v2"
+release_date = "2026-03-31"
+last_updated = "2026-03-31"
+attachment = true
+reasoning = false
+temperature = false
+tool_call = false
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 128_000
+output = 4_096
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]

providers/nvidia/models/nvidia/magpie-tts-zeroshot.toml

@@ -0,0 +1,21 @@
+name = "magpie-tts-zeroshot"
+description = "Speech generation model for controllable voice, narration, and audio delivery"
+release_date = "2025-05-22"
+last_updated = "2025-06-12"
+attachment = true
+reasoning = false
+temperature = false
+tool_call = false
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 0
+output = 4_096
+
+[modalities]
+input = ["text", "audio"]
+output = ["audio"]

providers/nvidia/models/nvidia/nemotron-3-content-safety.toml

@@ -0,0 +1,21 @@
+name = "nemotron-3-content-safety"
+base_model = "nvidia/nemotron-3-content-safety"
+release_date = "2026-04-16"
+last_updated = "2026-04-16"
+attachment = false
+reasoning = false
+temperature = false
+tool_call = false
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 128_000
+output = 4_096
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/nvidia/nemotron-3-nano-30b-a3b.toml

@@ -0,0 +1,25 @@
+name = "nemotron-3-nano-30b-a3b"
+base_model = "nvidia/nemotron-3-nano-30b-a3b"
+family = "nemotron"
+
+release_date = "2024-12"
+last_updated = "2024-12"
+attachment = false
+reasoning = true
+reasoning_options = [{ type = "toggle" }]
+temperature = true
+knowledge = "2024-09"
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 131_072
+output = 131_072
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/nvidia/nemotron-3-nano-omni-30b-a3b-reasoning.toml

@@ -0,0 +1,29 @@
+name = "Nemotron 3 Nano Omni"
+base_model = "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning"
+family = "nemotron"
+
+release_date = "2026-04-28"
+last_updated = "2026-04-28"
+attachment = true
+reasoning = true
+# NIM Chat schema: `reasoning_budget` is -1..32768 (default 16384); -1 means
+# unlimited/no budget enforcement, not reasoning off. `/think` and `/no_think`
+# go in a system message; video and audio requests should use `/no_think`.
+# https://docs.api.nvidia.com/nim/reference/nvidia-nemotron-3-nano-omni-30b-a3b-reasoning-infer
+reasoning_options = [{ type = "toggle" }, { type = "budget_tokens", min = -1, max = 32768 }]
+temperature = true
+tool_call = true
+structured_output = true
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 256_000
+output = 65_536
+
+[modalities]
+input = ["text", "image", "video", "audio"]
+output = ["text"]

providers/nvidia/models/nvidia/nemotron-3-super-120b-a12b.toml

@@ -0,0 +1,24 @@
+name = "Nemotron 3 Super"
+base_model = "nvidia/nemotron-3-super-120b-a12b"
+family = "nemotron"
+attachment = false
+reasoning = true
+reasoning_options = [{ type = "toggle" }]
+tool_call = true
+temperature = true
+knowledge = "2024-04"
+release_date = "2026-03-11"
+last_updated = "2026-03-11"
+open_weights = true
+
+[cost]
+input = 0.20
+output = 0.80
+
+[limit]
+context = 262144
+output = 262144
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/nvidia/nemotron-3-ultra-550b-a55b.toml

@@ -0,0 +1,12 @@
+base_model = "nvidia/nemotron-3-ultra-550b-a55b"
+reasoning_options = [{ type = "toggle" }]
+structured_output = true
+
+[cost]
+input = 0.50
+output = 2.50
+cache_read = 0.15
+
+[limit]
+context = 1_000_000
+output = 65_536

providers/nvidia/models/nvidia/nemotron-content-safety-reasoning-4b.toml

@@ -0,0 +1,22 @@
+name = "nemotron-content-safety-reasoning-4b"
+base_model = "nvidia/nemotron-content-safety-reasoning-4b"
+release_date = "2026-01-22"
+last_updated = "2026-01-22"
+attachment = false
+reasoning = true
+reasoning_options = []
+temperature = false
+tool_call = false
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 128_000
+output = 4_096
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/nvidia/nemotron-mini-4b-instruct.toml

@@ -0,0 +1,21 @@
+name = "nemotron-mini-4b-instruct"
+base_model = "nvidia/nemotron-mini-4b-instruct"
+release_date = "2024-08-21"
+last_updated = "2024-08-26"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 128_000
+output = 8_192
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/nvidia/nemotron-voicechat.toml

@@ -0,0 +1,21 @@
+name = "nemotron-voicechat"
+base_model = "nvidia/nemotron-voicechat"
+release_date = "2026-03-16"
+last_updated = "2026-03-16"
+attachment = true
+reasoning = false
+temperature = true
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 128_000
+output = 8_192
+
+[modalities]
+input = ["text", "audio"]
+output = ["text"]

providers/nvidia/models/nvidia/nv-embed-v1.toml

@@ -0,0 +1,21 @@
+name = "nv-embed-v1"
+description = "Embedding model for semantic search, retrieval, clustering, and ranking pipelines"
+release_date = "2024-06-07"
+last_updated = "2025-07-22"
+attachment = false
+reasoning = false
+temperature = false
+tool_call = false
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 32_768
+output = 2_048
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/nvidia/nv-embedcode-7b-v1.toml

@@ -0,0 +1,21 @@
+name = "nv-embedcode-7b-v1"
+description = "Nemotron model for efficient reasoning, coding, and specialized AI agents"
+release_date = "2025-03-17"
+last_updated = "2025-05-29"
+attachment = false
+reasoning = false
+temperature = false
+tool_call = false
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 32_768
+output = 2_048
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/nvidia/nvidia-nemotron-nano-9b-v2.toml

@@ -0,0 +1,25 @@
+name = "nvidia-nemotron-nano-9b-v2"
+base_model = "nvidia/nemotron-nano-9b-v2"
+family = "nemotron"
+
+release_date = "2025-08-18"
+last_updated = "2025-08-18"
+attachment = false
+reasoning = true
+reasoning_options = [{ type = "toggle" }]
+temperature = true
+knowledge = "2024-09"
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 131_072
+output = 131_072
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/nvidia/rerank-qa-mistral-4b.toml

@@ -0,0 +1,21 @@
+name = "rerank-qa-mistral-4b"
+description = "Reranking model for improving retrieval quality in search and recommendation systems"
+release_date = "2024-03-17"
+last_updated = "2025-01-17"
+attachment = false
+reasoning = false
+temperature = false
+tool_call = false
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 128_000
+output = 4_096
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/nvidia/riva-translate-4b-instruct-v1_1.toml

@@ -0,0 +1,21 @@
+name = "riva-translate-4b-instruct-v1_1"
+description = "Translation model for multilingual conversion, localization, and cross-language workflows"
+release_date = "2025-12-12"
+last_updated = "2025-12-12"
+attachment = false
+reasoning = false
+temperature = false
+tool_call = false
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 128_000
+output = 4_096
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/nvidia/sparsedrive.toml

@@ -0,0 +1,21 @@
+name = "sparsedrive"
+description = "Nemotron multimodal model for visual reasoning and agentic AI workflows"
+release_date = "2025-03-18"
+last_updated = "2025-07-20"
+attachment = true
+reasoning = false
+temperature = true
+tool_call = false
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 128_000
+output = 8_192
+
+[modalities]
+input = ["video"]
+output = ["text"]

providers/nvidia/models/nvidia/streampetr.toml

@@ -0,0 +1,21 @@
+name = "streampetr"
+description = "Nemotron multimodal model for visual reasoning and agentic AI workflows"
+release_date = "2025-11-13"
+last_updated = "2025-11-13"
+attachment = true
+reasoning = false
+temperature = true
+tool_call = false
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 128_000
+output = 8_192
+
+[modalities]
+input = ["video"]
+output = ["text"]

providers/nvidia/models/nvidia/studiovoice.toml

@@ -0,0 +1,21 @@
+name = "studiovoice"
+description = "Nemotron model for efficient reasoning, coding, and specialized AI agents"
+release_date = "2024-10-03"
+last_updated = "2025-06-13"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = false
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 128_000
+output = 8_192
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/nvidia/synthetic-video-detector.toml

@@ -0,0 +1,21 @@
+name = "synthetic-video-detector"
+description = "Video model for prompt-guided generation, editing, and motion workflows"
+release_date = "2026-04-16"
+last_updated = "2026-04-16"
+attachment = true
+reasoning = false
+temperature = true
+tool_call = false
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 0
+output = 4_096
+
+[modalities]
+input = ["video"]
+output = ["text"]

providers/nvidia/models/nvidia/usdcode.toml

@@ -0,0 +1,21 @@
+name = "usdcode"
+description = "Nemotron model for efficient reasoning, coding, and specialized AI agents"
+release_date = "2026-01-01"
+last_updated = "2026-01-01"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = false
+open_weights = false
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 128_000
+output = 4_096
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/nvidia/usdvalidate.toml

@@ -0,0 +1,21 @@
+name = "usdvalidate"
+description = "Nemotron model for efficient reasoning, coding, and specialized AI agents"
+release_date = "2024-07-24"
+last_updated = "2025-01-08"
+attachment = false
+reasoning = false
+temperature = false
+tool_call = false
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 0
+output = 4_096
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/openai/gpt-oss-120b.toml

@@ -0,0 +1,14 @@
+base_model = "openai/gpt-oss-120b"
+name = "GPT-OSS-120B"
+release_date = "2025-08-04"
+last_updated = "2025-08-14"
+knowledge = "2025-08"
+reasoning_options = [{ type = "effort", values = ["low", "medium", "high"] }]
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 128_000
+output = 8_192

providers/nvidia/models/openai/gpt-oss-20b.toml

@@ -0,0 +1,24 @@
+name = "GPT OSS 20B"
+description = "Open-weight GPT model for self-hosted reasoning and instruction-following workloads"
+family = "gpt-oss"
+release_date = "2025-08-05"
+last_updated = "2025-08-05"
+attachment = false
+reasoning = true
+reasoning_options = [{ type = "effort", values = ["low", "medium", "high"] }]
+temperature = true
+tool_call = true
+structured_output = true
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 131_072
+output = 32_768
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/openai/whisper-large-v3.toml

@@ -0,0 +1,23 @@
+name = "Whisper Large v3"
+description = "Speech transcription model for accurate audio-to-text and captioning workflows"
+family = "whisper"
+release_date = "2023-09-01"
+last_updated = "2025-09-05"
+attachment = false
+reasoning = false
+temperature = false
+knowledge = "2023-09"
+tool_call = false
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 0
+output = 4096
+
+[modalities]
+input = ["audio"]
+output = ["text"]

providers/nvidia/models/qwen/qwen-image-edit.toml

@@ -0,0 +1,23 @@
+name = "Qwen Image Edit"
+description = "Image model for prompt-driven generation, editing, and visual design workflows"
+family = "qwen"
+release_date = "2025-08-19"
+last_updated = "2025-08-19"
+attachment = true
+reasoning = false
+temperature = true
+tool_call = false
+structured_output = false
+open_weights = false
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 0
+output = 0
+
+[modalities]
+input = ["text", "image"]
+output = ["image"]

providers/nvidia/models/qwen/qwen-image.toml

@@ -0,0 +1,23 @@
+name = "Qwen Image"
+description = "Image model for prompt-driven generation, editing, and visual design workflows"
+family = "qwen"
+release_date = "2025-08-07"
+last_updated = "2025-08-07"
+attachment = true
+reasoning = false
+temperature = true
+tool_call = false
+structured_output = false
+open_weights = false
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 0
+output = 0
+
+[modalities]
+input = ["text", "image"]
+output = ["image"]

providers/nvidia/models/qwen/qwen2.5-coder-32b-instruct.toml

@@ -0,0 +1,22 @@
+name = "Qwen2.5 Coder 32b Instruct"
+description = "Qwen coding model for software agents, repository edits, and code reasoning"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = true
+structured_output = true
+release_date = "2024-11-06"
+last_updated = "2024-11-06"
+open_weights = true
+
+[cost]
+input = 0.00
+output = 0.00
+
+[limit]
+context = 128000
+output = 4096
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/qwen/qwen3-coder-480b-a35b-instruct.toml

@@ -0,0 +1,23 @@
+name = "Qwen3 Coder 480B A35B Instruct"
+description = "Qwen coding model for software agents, repository edits, and code reasoning"
+family = "qwen"
+release_date = "2025-07-23"
+last_updated = "2025-07-23"
+attachment = false
+reasoning = false
+temperature = true
+knowledge = "2025-04"
+tool_call = true
+open_weights = false
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 262_144
+output = 66_536
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/qwen/qwen3-next-80b-a3b-instruct.toml

@@ -0,0 +1,23 @@
+name = "Qwen3-Next-80B-A3B-Instruct"
+description = "Qwen instruction model for multilingual chat, reasoning, and tool use"
+family = "qwen"
+release_date = "2024-12-01"
+last_updated = "2025-09-05"
+attachment = false
+reasoning = false
+temperature = true
+knowledge = "2024-12"
+tool_call = true
+open_weights = false
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 262_144
+output = 16_384
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/qwen/qwen3.5-122b-a10b.toml

@@ -0,0 +1,26 @@
+name = "Qwen3.5 122B-A10B"
+description = "Qwen vision-language model for visual reasoning, documents, and agent tasks"
+family = "qwen"
+release_date = "2026-02-23"
+last_updated = "2026-02-23"
+attachment = true
+reasoning = true
+# NIM Chat schema: `chat_template_kwargs.enable_thinking = true|false`.
+# https://docs.api.nvidia.com/nim/reference/qwen-qwen3-5-122b-a10b-infer
+reasoning_options = [{ type = "toggle" }]
+temperature = true
+tool_call = true
+structured_output = true
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 262_144
+output = 65_536
+
+[modalities]
+input = ["text", "image", "video", "audio"]
+output = ["text"]

providers/nvidia/models/qwen/qwen3.5-397b-a17b.toml

@@ -0,0 +1,27 @@
+name = "Qwen3.5-397B-A17B"
+description = "Qwen vision-language model for visual reasoning, documents, and agent tasks"
+family = "qwen"
+release_date = "2026-02-16"
+last_updated = "2026-02-16"
+attachment = true
+reasoning = true
+# NIM Chat schema: `chat_template_kwargs.enable_thinking = true|false`.
+# https://docs.api.nvidia.com/nim/reference/qwen-qwen3-5-397b-a17b-infer
+reasoning_options = [{ type = "toggle" }]
+temperature = true
+knowledge = "2026-01"
+tool_call = true
+structured_output = true
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 262_144
+output = 8_192
+
+[modalities]
+input = ["text", "image"] # video is natively supported by Qwen3.5 but not available on NVIDIA NIM
+output = ["text"]

providers/nvidia/models/sarvamai/sarvam-m.toml

@@ -0,0 +1,21 @@
+name = "sarvam-m"
+description = "Efficient Indian-language reasoning model for chat, coding, and multilingual work"
+release_date = "2025-07-25"
+last_updated = "2025-07-25"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 128_000
+output = 8_192
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/stepfun-ai/step-3.5-flash.toml

@@ -0,0 +1,22 @@
+name = "Step 3.5 Flash"
+description = "StepFun flash model for efficient multimodal reasoning, coding, and tool use"
+release_date = "2026-02-02"
+last_updated = "2026-02-02"
+attachment = false
+reasoning = true
+reasoning_options = [{ type = "effort", values = ["low", "medium", "high"] }]
+temperature = true
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 256_000
+output = 16_384
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/stepfun-ai/step-3.7-flash.toml

@@ -0,0 +1,22 @@
+name = "Step 3.7 Flash"
+description = "StepFun flash model for efficient multimodal reasoning, coding, and tool use"
+release_date = "2026-05-28"
+last_updated = "2026-05-28"
+attachment = true
+reasoning = true
+reasoning_options = [{ type = "effort", values = ["minimal", "low", "medium", "high", "xhigh", "max"] }]
+temperature = true
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 256_000
+output = 16_384
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]

providers/nvidia/models/upstage/solar-10_7b-instruct.toml

@@ -0,0 +1,21 @@
+name = "solar-10.7b-instruct"
+description = "Open-weight instruction model for adaptable chat and self-hosted production workloads"
+release_date = "2024-06-05"
+last_updated = "2025-04-10"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 128_000
+output = 8_192
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/nvidia/models/z-ai/glm-5.2.toml

@@ -0,0 +1,12 @@
+# NIM Chat schema: `chat_template_kwargs.enable_thinking = true|false`.
+# https://docs.api.nvidia.com/nim/reference/z-ai-glm-5.2-infer
+base_model = "zhipuai/glm-5.2"
+
+reasoning_options = [{ type = "toggle" }]
+
+[interleaved]
+field = "reasoning_content"
+
+[cost]
+input = 0.0
+output = 0.0