Cohere 'Too Many Tokens' Context Length Error

How to fix it

Try these in order — the most common fix is first.

1. 1. Trim what you send

   Copy steps
   1. Summarize or drop older turns
   2. Send only top reranked passages instead of all retrieved text
   Did this fix it? Yes No

2. 2. Use rerank before generation

   Copy steps
   1. Rerank retrieved chunks
   2. Pass the top few to the Chat call
   Did this fix it? Yes No

3. 3. Choose a longer-context model

   Copy steps
   1. Check the model's context length in the models docs
   2. Switch to a current Command A model if needed
   Did this fix it? Yes No

Symptoms

A 400 error says the prompt or inputs are too long for the model.

Why it happens

• Long chat history sent each turn
• Large RAG documents pasted into the prompt
• Using a model with a smaller context window

This guide also covers

• cohere too many tokens
• cohere context length exceeded
• cohere prompt too long

The same problem in other apps

All not loading, error messages & connection guides →

• Claude API 529 Overloaded Error (Anthropic Claude)

Other Cohere errors

• Cohere API 429 Too Many Requests Error
• Cohere Trial Key Limits and 'Not for Production' Error
• Cohere 401 Invalid API Token Error
• Cohere Model Not Found Error (Deprecated Models)
• Cohere Embed Input Too Long or Too Many Texts Error
• Cohere Rerank API Errors and Truncated Documents
• Cohere API 500 / 503 / 504 Server Errors
• How to Cancel Cohere (Delete Account and Stop Billing)