Feat/update nvidia models

[b3nw]

add some missing models from nvidia nim.

[Copilot]

Pull request overview

This PR adds 43 missing model configurations for NVIDIA NIM, expanding support across multiple model providers including Qwen, NVIDIA, Mistral AI, Microsoft Phi, Meta Llama, Google Gemma, and DeepSeek AI. The configurations follow a standardized TOML format defining model capabilities, costs, limits, and modalities.

Key Changes

• Added reasoning-capable models: Qwq-32b, DeepSeek R1 variants
• Added vision-capable models: Phi-3 variants, Llama 3.2/4 variants, Gemma 3 variants
• Added code-specialized models: Qwen2.5 Coder variants, CodeGemma, Codestral variants

Reviewed changes

Copilot reviewed 42 out of 42 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
qwen/qwq-32b.toml	Adds Qwq 32b reasoning model with 128k context
qwen/qwen2.5-coder-7b-instruct.toml	Adds 7B parameter coding model with tool calling
qwen/qwen2.5-coder-32b-instruct.toml	Adds 32B parameter coding model with tool calling
nvidia/usdcode-llama-3.1-70b-instruct.toml	Adds USD-specialized coding model
nvidia/nemotron-4-340b-instruct.toml	Adds large 340B parameter instruct model
nvidia/llama3-chatqa-1.5-70b.toml	Adds conversation-focused model
nvidia/llama-3.3-nemotron-super-49b-v1.toml	Adds Nemotron super model v1
nvidia/llama-3.3-nemotron-super-49b-v1.5.toml	Adds Nemotron super model v1.5
nvidia/llama-3.1-nemotron-70b-instruct.toml	Adds 70B Nemotron instruct model
nvidia/llama-3.1-nemotron-51b-instruct.toml	Adds 51B Nemotron instruct model
mistralai/mistral-small-3.1-24b-instruct-2503.toml	Adds Mistral Small 3.1 model
mistralai/mistral-large-2-instruct.toml	Adds Mistral Large 2 model
mistralai/mamba-codestral-7b-v0.1.toml	Adds Mamba-based coding model
mistralai/codestral-22b-instruct-v0.1.toml	Adds Codestral 22B coding model
microsoft/phi-3.5-vision-instruct.toml	Adds Phi 3.5 vision-capable model
microsoft/phi-3.5-moe-instruct.toml	Adds Phi 3.5 MoE model
microsoft/phi-3-vision-128k-instruct.toml	Adds Phi 3 vision model with 128k context
microsoft/phi-3-small-8k-instruct.toml	Adds Phi 3 small model with 8k context
microsoft/phi-3-small-128k-instruct.toml	Adds Phi 3 small model with 128k context
microsoft/phi-3-medium-4k-instruct.toml	Adds Phi 3 medium model with 4k context
microsoft/phi-3-medium-128k-instruct.toml	Adds Phi 3 medium model with 128k context
meta/llama3-8b-instruct.toml	Adds Llama 3 8B instruct model
meta/llama3-70b-instruct.toml	Adds Llama 3 70B instruct model
meta/llama-4-scout-17b-16e-instruct.toml	Adds Llama 4 Scout vision model
meta/llama-4-maverick-17b-128e-instruct.toml	Adds Llama 4 Maverick vision model
meta/llama-3.3-70b-instruct.toml	Adds Llama 3.3 70B instruct model
meta/llama-3.2-1b-instruct.toml	Adds Llama 3.2 1B instruct model
meta/llama-3.2-11b-vision-instruct.toml	Adds Llama 3.2 11B vision model
meta/llama-3.1-70b-instruct.toml	Adds Llama 3.1 70B instruct model
meta/llama-3.1-405b-instruct.toml	Adds Llama 3.1 405B instruct model
meta/codellama-70b.toml	Adds CodeLlama 70B coding model
google/gemma-3n-e4b-it.toml	Adds Gemma 3n E4b vision model
google/gemma-3n-e2b-it.toml	Adds Gemma 3n E2b vision model
google/gemma-3-1b-it.toml	Adds Gemma 3 1B vision model
google/gemma-3-12b-it.toml	Adds Gemma 3 12B model
google/gemma-2-2b-it.toml	Adds Gemma 2 2B model
google/gemma-2-27b-it.toml	Adds Gemma 2 27B model
google/codegemma-7b.toml	Adds CodeGemma 7B coding model
google/codegemma-1.1-7b.toml	Adds CodeGemma 1.1 7B coding model
deepseek-ai/deepseek-r1.toml	Adds DeepSeek R1 reasoning model
deepseek-ai/deepseek-r1-0528.toml	Adds DeepSeek R1 0528 variant
deepseek-ai/deepseek-coder-6.7b-instruct.toml	Adds DeepSeek Coder 6.7B model

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

The model claims to support both text and image input modalities but has attachment set to false. Models that support image input typically require attachment to be set to true to handle image uploads. Please verify this configuration is correct.

[b3nw]

fp, attachment = true in the cited code