-
What happened?Claude 3.7 Sonnet incorrectly identifies itself as "Claude 3 Opus" in LibreChat responses, despite correctly identifying as Claude 3.7 Sonnet in the official claude.ai interface. This is particularly concerning since Claude 3 Opus is a more expensive model. Version Informationghcr.io/danny-avila/librechat-dev latest 7768baf431da 3 hours ago 874MB Steps to ReproduceStep 1: Configure LibreChat to use Claude 3.7 Sonnet What browsers are you seeing the problem on?Chrome, Firefox Relevant log outputI'm Claude, an AI assistant created by Anthropic. Specifically, I'm Claude 3 Opus, which is Anthropic's most capable I'm Claude, an AI assistant created by Anthropic. Specifically, I'm Claude 3 Opus, Anthropic's most capable model. ScreenshotsCode of Conduct
|
Beta Was this translation helpful? Give feedback.
Replies: 2 comments
-
This is not a bug with LibreChat. OpenRouter serving the model via Bedrock has the same response. Recommend this be closed. |
Beta Was this translation helpful? Give feedback.
-
This isn't a LibreChat issue. Models like Claude don't inherently know their own name - they identify based on what they're told in the system prompt, and LibreChat doesn't. Newer Claude models are probably fine-tuned from Claude 3 Opus outputs, so they default to calling themselves "Opus" unless instructed otherwise. This should be closed |
Beta Was this translation helpful? Give feedback.
This isn't a LibreChat issue. Models like Claude don't inherently know their own name - they identify based on what they're told in the system prompt, and LibreChat doesn't.
Newer Claude models are probably fine-tuned from Claude 3 Opus outputs, so they default to calling themselves "Opus" unless instructed otherwise.
The official app likely includes the correct model name in its instructions
This should be closed