minimax-m3 on nim provider is missing #34177

Activity

1. github-actions

   assigned

   jlongster

   on Jun 27, 2026
2. github-actions commented on Jun 27, 2026

   

   github-actionsBot

   on Jun 27, 2026 – with GitHub Actions

   Contributor

   More actions

   This issue might be a duplicate of existing issues. Please check:

   • [FEATURE]: Add MiniMax M2.5 to NVIDIA NIM provider #14726: Adding MiniMax M2.5 to NVIDIA NIM provider (closed, same pattern for an earlier MiniMax version)
   • Add support for minimax-m2.7 model via Nvidia NIM connection #22427: Add support for minimax-m2.7 model via NVIDIA NIM (closed, same pattern for M2.7)
   • Nvidea nim added deepseek v4 flash and pro but they are not showing up i tried updating and reloading #24495: New NIM models not showing up after update (closed, same symptom)
   • Kimi 2.5 Reasoning Cutoff/Missing when using NVIDIA API #13515: Kimi 2.5 reasoning cutoff/missing when using NVIDIA API (open, directly related to the Kimi thinking issue you mentioned)
3. kirazizi commented on Jun 27, 2026

   

   kirazizi

   on Jun 27, 2026

   Last edited by kirazizi

   More actions

   1. Edit config:
   nano ~/.config/opencode/opencode.jsonc
   
   2. Use this content:
   {
     "$schema": "https://opencode.ai/config.json",
     "provider": {
       "nvidia": {
         "options": {
           "baseURL": "https://integrate.api.nvidia.com/v1",
           "apiKey": "YOUR_NVIDIA_API_KEY"
         },
         "models": {
           "minimaxai/minimax-m3": {
             "name": "MiniMax-M3"
           }
         }
       }
     }
   }
   
   3. Get your NVIDIA API key from: https://ngc.nvidia.com/setup/api-key
   4. Replace YOUR_NVIDIA_API_KEY in the config with your actual key
   5. Restart opencode — that's it!
   
4. clarkarch commented on Jun 28, 2026

   

   clarkarch

   on Jun 28, 2026

   Author

   More actions

   1. Edit config:
   nano ~/.config/opencode/opencode.jsonc
   
   2. Use this content:
   {
     "$schema": "https://opencode.ai/config.json",
     "provider": {
       "nvidia": {
         "options": {
           "baseURL": "https://integrate.api.nvidia.com/v1",
           "apiKey": "YOUR_NVIDIA_API_KEY"
         },
         "models": {
           "minimaxai/minimax-m3": {
             "name": "MiniMax-M3"
           }
         }
       }
     }
   }
   
   3. Get your NVIDIA API key from: https://ngc.nvidia.com/setup/api-key
   4. Replace YOUR_NVIDIA_API_KEY in the config with your actual key
   5. Restart opencode — that's it!
   

   how do i perform extra body parameters stuff? i wanna turn on thinking
5. rekram1-node commented on Jul 13, 2026

   

   rekram1-node

   on Jul 13, 2026

   Collaborator

   More actions

   Fixed in models.dev by anomalyco/models.dev#2621. The NVIDIA catalog now includes minimaxai/minimax-m3 with its supported reasoning toggle. The NVIDIA Kimi K2.6 entry also carries reasoning metadata. Closing as completed; refresh the model catalog or use an OpenCode release containing the updated snapshot.
6. rekram1-node

   closed this as completedon Jul 13, 2026