This message was deleted.

On the openhands cloud the LM calls are now going through the All Hands account. Right now the model is claude-3.7-sonnet

Ok, but why is it giving some incorrect information when I asked about which model is being used? or I am understanding it correctly?

Where did you get that information? By asking OpenHands? Or is it on a doc page?

by asking OpenHands, here is the full response ---

Based on my investigation, I can now provide you with information about the model we're using:

Based on my investigation of the environment, I can tell you that the default model being used in this OpenHands environment is Claude 3.7 Sonnet.

This is evident from the LLM configuration file at /openhands/code/openhands/core/config/llm_config.py, where on line 49 we can see:

model: str = Field(default='claude-3-7-sonnet-20250219')
This indicates that the default model is Claude 3.7 Sonnet, specifically the version from February 19, 2025 (as indicated by the date code in the model name).

The environment also has support for other models, as we can see from the installed libraries:

OpenAI (version 1.66.3)
Anthropic (version 0.49.0)
Google's generative AI libraries
The system is designed to work with various models including:

Claude models (3.7 Sonnet, 3.5 Sonnet, 3.5 Haiku, etc.)
OpenAI models (GPT-4o, GPT-4o-mini, etc.)
Anthropic's O1 and O3 models
But the default model configured is Claude 3.7 Sonnet.


I believe that the task was completed successfully.

Hmm, that may be a hallucination 😕

ohh ok

hah that actually looks like it legit grabbed info from our OSS repo. I think it's all accurate (for the OSS)

But “anthropic o1” is not accurate

😂 I missed that