Fix Bedrock HTTP 400 for gpt-5.6-sol by using auto reasoning summary - #111
Merged
Merged
Conversation
Bedrock rejects explicit reasoning.summary levels ("concise"/"detailed")
for some models (e.g. openai.gpt-5.6-sol), failing requests with HTTP
400. "auto" is the OpenAI Responses default and is accepted; Mantle
does not return summary text either way, so behavior is unchanged.
wch
enabled auto-merge (squash)
September 10, 2026 21:08
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
When a host application chats with a GPT 5.6-series model through Amazon Bedrock's OpenAI Responses route (for example,
openai.gpt-5.6-sol), every request fails withHTTP 400: Unsupported parameter: 'reasoning.summary' is not supported with the 'openai.gpt-5.6-sol' model.This started suddenly with no change on the client's side: Bedrock began rejecting the explicit reasoning-summary level (concise/detailed) the client sends on that route.This PR sends
auto— the OpenAI Responses API's own default — instead of an explicit level. Bedrock acceptsautofor these models, and the route never returned reasoning-summary text either way, so the only visible change is that requests succeed again.Only the Bedrock OpenAI Responses route changes. The Bedrock chat route, Anthropic and Converse routes, and every other provider are untouched. If Bedrock's validation later tightens further and rejects the field entirely, omitting it is the remaining fallback.
Verification
The Mantle routing test now asserts the
autolevel, and the change was confirmed against the liveopenai.gpt-5.6-solmodel: requests complete and stream a response instead of returning HTTP 400.