orcs-oss/orcs-code
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat: add NVIDIA NIM and MiniMax provider support (#552)

Browse Source 
* feat: add NVIDIA NIM and MiniMax provider support

- Add nvidia-nim and minimax to --provider CLI flag
- Add model discovery for NVIDIA NIM (160+ models) and MiniMax
- Update /model picker to show provider-specific models
- Fix provider detection in startup banner
- Update .env.example with new provider options

Supported providers:
- NVIDIA NIM: https://integrate.api.nvidia.com/v1
- MiniMax: https://api.minimax.io/v1

* fix: resolve conflict in StartupScreen (keep NVIDIA/MiniMax + add Codex detection)

* fix: resolve providerProfile conflict (add imports from main, keep NVIDIA/MiniMax)

* fix: revert providerSecrets to match main (NVIDIA/MiniMax handled elsewhere)

* fix: add context window entries for NVIDIA NIM and new MiniMax models

* fix: use GLM-5 as NVIDIA NIM default and MiniMax-M2.5 for consistency

* fix: address remaining review items - add GLM/Kimi context entries, max output tokens, fix .env.example, revert to Nemotron default

* fix: filter NVIDIA NIM picker to chat/instruct models only, set provider-specific API keys from saved profiles

* chore: add more NVIDIA NIM context window entries for popular models

* fix: address remaining non-blocking items - fix base model, clear provider API keys on profile switch
This commit is contained in:
ArkhAngelLifeJiggy
2026-04-15 13:26:13 +01:00
committed by GitHub
parent 6b2121da12
commit 51191d6132
15 changed files with 628 additions and 70 deletions
Download Patch File Download Diff File Expand all files Collapse all files

95
src/utils/model/openaiContextWindows.ts
View File

@@ -104,6 +104,57 @@ const OPENAI_CONTEXT_WINDOWS: Record<string, number> = {
'devstral-latest': 256_000,
'ministral-3b-latest': 256_000,
// NVIDIA NIM - popular models
'nvidia/llama-3.1-nemotron-70b-instruct': 128_000,
'nvidia/llama-3.1-nemotron-ultra-253b-v1': 128_000,
'nvidia/nemotron-mini-4b-instruct': 32_768,
'meta/llama-3.1-405b-instruct': 128_000,
'meta/llama-3.1-70b-instruct': 128_000,
'meta/llama-3.1-8b-instruct': 128_000,
'meta/llama-3.2-90b-instruct': 128_000,
'meta/llama-3.2-1b-instruct': 128_000,
'meta/llama-3.2-3b-instruct': 128_000,
'meta/llama-3.3-70b-instruct': 128_000,
// Google Gemma via NVIDIA NIM
'google/gemma-2-27b-it': 8_192,
'google/gemma-2-9b-it': 8_192,
'google/gemma-3-27b-it': 131_072,
'google/gemma-3-12b-it': 131_072,
'google/gemma-3-4b-it': 131_072,
// DeepSeek via NVIDIA NIM
'deepseek-ai/deepseek-r1': 128_000,
'deepseek-ai/deepseek-v3': 128_000,
'deepseek-ai/deepseek-v3.2': 128_000,
// Qwen via NVIDIA NIM
'qwen/qwen3-32b': 128_000,
'qwen/qwen3-8b': 128_000,
'qwen/qwen2.5-7b-instruct': 32_768,
// Mistral via NVIDIA NIM
'mistralai/mistral-large-3-675b-instruct-2512': 256_000,
'mistralai/mistral-large-2-instruct': 256_000,
'mistralai/mistral-small-3.1-24b-instruct-2503': 32_768,
'mistralai/mixtral-8x7b-instruct-v0.1': 32_768,
// Microsoft Phi via NVIDIA NIM
'microsoft/phi-4-mini-instruct': 16_384,
'microsoft/phi-3.5-mini-instruct': 16_384,
'microsoft/phi-3-mini-128k-instruct': 128_000,
// IBM Granite via NVIDIA NIM
'ibm/granite-3.3-8b-instruct': 8_192,
'ibm/granite-8b-code-instruct': 8_192,
// GLM models via NVIDIA NIM
'z-ai/glm5': 200_000,
'z-ai/glm4.7': 128_000,
// Kimi models via NVIDIA NIM
'moonshotai/kimi-k2.5': 200_000,
'moonshotai/kimi-k2-instruct': 128_000,
// DBRX via NVIDIA NIM
'databricks/dbrx-instruct': 131_072,
// Jamba via NVIDIA NIM
'ai21labs/jamba-1.5-large-instruct': 256_000,
'ai21labs/jamba-1.5-mini-instruct': 256_000,
// Yi via NVIDIA NIM
'01-ai/yi-large': 32_768,
// MiniMax (all M2.x variants share 204,800 context, 131,072 max output)
'MiniMax-M2.7': 204_800,
'MiniMax-M2.7-highspeed': 204_800,
@@ -118,6 +169,13 @@ const OPENAI_CONTEXT_WINDOWS: Record<string, number> = {
'minimax-m2.1': 204_800,
'minimax-m2.1-highspeed': 204_800,
// MiniMax new models
'MiniMax-Text-01': 524_288,
'MiniMax-Text-01-Preview': 262_144,
'MiniMax-Vision-01': 32_768,
'MiniMax-Vision-01-Fast': 16_384,
'MiniMax-M2': 204_800,
// Google (via OpenRouter)
'google/gemini-2.0-flash':1_048_576,
'google/gemini-2.5-pro': 1_048_576,
@@ -246,6 +304,12 @@ const OPENAI_MAX_OUTPUT_TOKENS: Record<string, number> = {
'minimax-m2.5-highspeed': 131_072,
'minimax-m2.1': 131_072,
'minimax-m2.1-highspeed': 131_072,
// New MiniMax models
'MiniMax-M2': 131_072,
'MiniMax-Text-01': 65_536,
'MiniMax-Text-01-Preview': 65_536,
'MiniMax-Vision-01': 16_384,
'MiniMax-Vision-01-Fast': 16_384,
// Google (via OpenRouter)
'google/gemini-2.0-flash': 8_192,
@@ -266,11 +330,32 @@ const OPENAI_MAX_OUTPUT_TOKENS: Record<string, number> = {
'deepseek-r1:14b': 8_192,
'mistral:7b': 4_096,
'phi4:14b': 4_096,
'gemma2:27b': 4_096,
'codellama:13b': 4_096,
'llama3.2:1b': 4_096,
'qwen3:8b': 8_192,
'codestral': 8_192,
// NVIDIA NIM models
'nvidia/llama-3.1-nemotron-70b-instruct': 32_768,
'nvidia/nemotron-mini-4b-instruct': 8_192,
'meta/llama-3.1-405b-instruct': 32_768,
'meta/llama-3.1-70b-instruct': 32_768,
'meta/llama-3.2-90b-instruct': 32_768,
'meta/llama-3.3-70b-instruct': 32_768,
'google/gemma-2-27b-it': 4_096,
'google/gemma-3-27b-it': 16_384,
'google/gemma-3-12b-it': 16_384,
'deepseek-ai/deepseek-r1': 32_768,
'deepseek-ai/deepseek-v3': 32_768,
'deepseek-ai/deepseek-v3.2': 32_768,
'qwen/qwen3-32b': 32_768,
'qwen/qwen2.5-7b-instruct': 8_192,
'mistralai/mistral-large-3-675b-instruct-2512': 32_768,
'mistralai/mixtral-8x7b-instruct-v0.1': 8_192,
'microsoft/phi-4-mini-instruct': 4_096,
'microsoft/phi-3.5-mini-instruct': 4_096,
'ibm/granite-3.3-8b-instruct': 4_096,
'z-ai/glm5': 32_768,
'moonshotai/kimi-k2.5': 32_768,
'databricks/dbrx-instruct': 32_768,
'ai21labs/jamba-1.5-large-instruct': 32_768,
'01-ai/yi-large': 8_192,
}
function lookupByModel<T>(table: Record<string, T>, model: string): T | undefined {