NVIDIA / NVIDIA/NemoClaw

[All Platforms][Onboard] NVIDIA Endpoints wizard lists an unavailable MiniMax M3 model

Open
#11,364 0 comments 0 reactions 0 assignees View on GitHub
area: onboarding integration: hermes integration: openclaw NV QA provider: nvidia
Dominant language
TypeScript
Stars
22.5k
Forks
3.1k
Avg merge
1d 1h
Merged PRs (30d)
715

Description

## Description

NemoClaw v0.0.121 presents MiniMax M3 as a selectable NVIDIA Endpoints model even though that model is absent from the live NVIDIA Endpoints model catalog. The mismatch is reproduced in both the OpenClaw and Hermes onboarding flows.

Platform scope: Reproduced on macOS arm64, WSL2 x86_64, and WSL2 arm64.

Regression: Unknown — earlier versions were not tested against the same live catalog state.

OpenShell issue: No.

## Environment

```text
Device: macOS arm64 and WSL2 x86_64/arm64 test hosts
OS: macOS 26.5 and WSL2 Linux
Architecture: arm64 and x86_64
Node.js: v22.23.2
npm: 10.9.8
Docker: Docker Engine 29.5.2 on macOS; Docker Desktop integration on WSL2
OpenShell CLI: openshell 0.0.106
NemoClaw: nemoclaw v0.0.121
OpenClaw: Not required for the catalog comparison; both OpenClaw and Hermes onboarding menus are affected
```

## Steps to Reproduce

1. Install NemoClaw v0.0.121.
2. Start either the OpenClaw or Hermes onboarding flow.
3. Select NVIDIA Endpoints as the inference provider and inspect the offered model list.
4. Query the live NVIDIA Endpoints models API with valid credentials.
5. Compare the wizard model identifiers with the live catalog identifiers.

## Expected Result

Every NVIDIA Endpoints model offered by the onboarding wizard is present in the live models API response.

## Actual Result

The wizard offers `minimaxai/minimax-m3` (MiniMax M3), but the live models API response does not contain `minimaxai/minimax-m3`. The unavailable selection is exposed in both the OpenClaw and Hermes onboarding flows.

## Logs

```text
Wizard-only model detected: minimaxai/minimax-m3
Live catalog membership: absent
Affected release: v0.0.121
```

Contributor guide

Open the contributing guide

Research direction

Locate the NVIDIA Endpoints model lists used by the OpenClaw and Hermes onboarding flows, then compare them with the live NVIDIA Endpoints models API using valid credentials. Verify the wizard no longer offers minimaxai/minimax-m3 and that every offered model appears in the live catalog.

Written by the indexing model from the issue text.

Assessment

Tech stack
typescript
Domain
ai, cli
Issue type
Bug
Difficulty
3/5
Estimated time
1-2 days
Activity status
Active
Clarity
Mostly clear
Newbie friendliness
68/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.