holetron-lab/fleet-memory
1
0
Fork 0
Code Releases Packages Activity Actions

fix: use max_tokens for OpenAI-compatible endpoints with custom base URL (#858)

Browse source 
Mistral (and several other providers) reject 'max_completion_tokens' with a 422
because they haven't adopted the newer OpenAI parameter name. When the openai
provider is configured with a custom base_url (e.g. Mistral, Together AI),
fall back to the widely-supported 'max_tokens' parameter.

Native OpenAI (no custom base_url) and Groq still use 'max_completion_tokens'.

Fixes #852
This commit is contained in:
Octopus 2026-04-07 02:13:08 -05:00 committed by GitHub
parent cd4b3e96e2
commit cd99eef4c5
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
1 changed files with 19 additions and 4 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

23
hindsight-api-slim/hindsight_api/engine/providers/openai_compatible_llm.py
View file

@ -191,6 +191,23 @@ class OpenAICompatibleLLM(LLMInterface):
return None return None
def _max_tokens_param_name(self) -> str:
"""Return the correct parameter name for limiting response tokens.
Native OpenAI and Groq accept 'max_completion_tokens'. Mistral and other
OpenAI-compatible endpoints that haven't adopted the newer parameter name
require 'max_tokens'. Using a custom base_url with the openai provider
signals a third-party compatible API, so fall back to 'max_tokens'.
"""
# Native OpenAI (no custom base URL) and Groq use max_completion_tokens
if self.provider == "groq":
return "max_completion_tokens"
if self.provider == "openai" and not self.base_url:
return "max_completion_tokens"
# openai with custom base_url, ollama, lmstudio, minimax, volcano —
# use the widely-supported max_tokens
return "max_tokens"
async def call( async def call(
self, self,
messages: list[dict[str, str]], messages: list[dict[str, str]],
@ -263,9 +280,7 @@ class OpenAICompatibleLLM(LLMInterface):
# For reasoning models, enforce minimum to ensure space for reasoning + output # For reasoning models, enforce minimum to ensure space for reasoning + output
if is_reasoning_model and max_completion_tokens < 16000: if is_reasoning_model and max_completion_tokens < 16000:
max_completion_tokens = 16000 max_completion_tokens = 16000
call_params["max_completion_tokens"] = max_completion_tokens call_params[self._max_tokens_param_name()] = max_completion_tokens
# Temperature - reasoning models don't support custom temperature
if temperature is not None and not is_reasoning_model: if temperature is not None and not is_reasoning_model:
# MiniMax requires temperature in (0.0, 1.0] — clamp accordingly # MiniMax requires temperature in (0.0, 1.0] — clamp accordingly
if self.provider == "minimax": if self.provider == "minimax":
@ -577,7 +592,7 @@ class OpenAICompatibleLLM(LLMInterface):
} }
if max_completion_tokens is not None: if max_completion_tokens is not None:
call_params["max_completion_tokens"] = max_completion_tokens call_params[self._max_tokens_param_name()] = max_completion_tokens
if temperature is not None: if temperature is not None:
# MiniMax requires temperature in (0.0, 1.0] — clamp accordingly # MiniMax requires temperature in (0.0, 1.0] — clamp accordingly
if self.provider == "minimax": if self.provider == "minimax":