openclaw - 💡(How to fix) Fix [Bug]: ollama cloud models are getting loaded with wrong context limit [1 comments, 2 participants]

Bug type

Regression (worked before, now fails)

Beta release blocker

No

Summary

Using 'openclaw models list --all --provider ollama' shows incorrect context limits.

Steps to reproduce

1. openclaw models list --all --provider ollama

For example nemotron-3-super shows context as 125k, whereas using the ollama.com api it shows up as 262144

Expected behavior

The context limits shown by openclaw models list --all --provider ollama should be the same as this command. $ curl -s https://ollama.com/api/show -H "Content-Type: application/json" -d '{"name":"nemotron-3-super"}' | jq { "details": { "parent_model": "nemotron-3-super", "format": "", "family": "nemotron_h_moe", "families": null, "parameter_size": "120000000000", "quantization_level": "NVFP4" }, "model_info": { "general.architecture": "nemotron_h_moe", "general.parameter_count": 120000000000, "nemotron_h_moe.context_length": 262144, "nemotron_h_moe.embedding_length": 4096 }, "capabilities": [ "completion", "thinking", "tools" ], "modified_at": "2026-03-11T00:00:00Z" }

Actual behavior

For example nemotron-3-super shows context as 125k, whereas using the ollama.com api it shows up as 262144

OpenClaw version

2026.4.21

Operating system

openclaw:latest

Install method

docker

Model

n/a/

Provider / routing chain

n/a

Additional provider/model setup details

No response

Logs, screenshots, and evidence

Impact and severity

No response

Additional information

No response