charieswei
/
TradingAgents
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Skip reasoning_effort entirely since all agents use function tools 

Per issue #403 analysis: the real problem is that reasoning_effort
with function tools requires the /v1/responses endpoint, but
LangChain's ChatOpenAI only supports /v1/chat/completions. This
affects all models (not just gpt-5.x vs o-series), so the safest
fix is to skip reasoning_effort and log an info message explaining
why.

Once LangChain adds /v1/responses support, this can be re-enabled.
This commit is contained in:
Charlie Tonneslan 2026-03-22 15:18:44 -04:00
parent 19dae36f58
commit 98330aa302
1 changed files with 14 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
tradingagents/graph/trading_graph.py
View File

@ -158,10 +158,20 @@ class TradingAgentsGraph:
elif provider == "openai": elif provider == "openai":
reasoning_effort = self.config.get("openai_reasoning_effort") reasoning_effort = self.config.get("openai_reasoning_effort")
if reasoning_effort: if reasoning_effort:
# reasoning_effort is only supported by o-series models # reasoning_effort + function tools is not supported on the
# (o1, o3, o3-mini, o4-mini, etc.), not by gpt-* models # /v1/chat/completions endpoint for some models (e.g. gpt-5.x).
if model_name.startswith("o"): # LangChain's ChatOpenAI only uses chat/completions, so we
kwargs["reasoning_effort"] = reasoning_effort # skip reasoning_effort for agents that use tools to avoid
# 400 errors. This is a known LangChain limitation.
# See: https://github.com/TauricResearch/TradingAgents/issues/403
import logging
logging.getLogger(__name__).info(
"Skipping reasoning_effort='%s' for model '%s' because "
"TradingAgents agents use function tools, which are "
"incompatible with reasoning_effort on the chat/completions "
"endpoint. See issue #403.",
reasoning_effort, model_name,
)
return kwargs return kwargs