claude-haiku-4-5-20251001-thinking
Creates a model response for the given chat conversation.
Sampling with Thinking: Omit temperature and top_k for reliable requests. The unchanged default temperature: 1 is accepted, and top_p supports values from 0.95 to 1. Other values return a 400 error while Thinking is enabled.
Authorizations
Bearer authentication header of the form Bearer <token>, where <token> is your auth token.
Body
Model ID
claude-haiku-4-5-20251001-thinking "claude-haiku-4-5-20251001-thinking"
A list of messages comprising the conversation so far.
1The maximum number of tokens to generate.
1 <= x <= 8192Keep the default value 1 or omit this field. Non-default values return a 400 error while Thinking is enabled.
1 <= x <= 11
Nucleus sampling threshold. While Thinking is enabled, values below 0.95 return a 400 error.
0.95 <= x <= 1Penalizes repeated tokens based on their frequency in the text so far.
-2 <= x <= 2Penalizes tokens that have already appeared in the text.
-2 <= x <= 2If true, stream partial message deltas using SSE.
Sequences where the model will stop generating further tokens.