Repository navigation
Error: 400 Missing required parameter: 'messages[2].reasoning_content'. on any Beam request #1073
Description
Summary
any request returns this on Beam
❯ bro?
❯ Error: 400 Missing required parameter: 'messages[2].reasoning_content'. This model needs the reasoning of
every earlier assistant message: send back the 'reasoning_content' returned with that message. If 're...
Type "continue" to try again. If the issue persists, contact support: https://commandcode.ai/discord
Trace ID: c4feb3d694a896a0b8f8d7d16ae4b68d
Expected Behavior
it should work
Actual Behavior
it doesn't work
Steps to reproduce the issue
just select beam model
Command Code Version
1.88.0
Operating System
macOS
Terminal/IDE
No response
Shell
No response
Session file (optional)
No response
Fix prompt (optional)
No response
Additional context
No response
Reacted by Marcus2Pierce
Activity
huizixin commentedon Nov 4, 2026
Adding another data point — this is reproducible on the hosted
reflection/beam:freemodel (via vercel-ai-gateway), not only BYOK/Beam.Environment:
reflection/beam:free(provider: vercel-ai-gateway),reasoningEffort: max400 Missing required parameter: 'messages[47].reasoning_content'. This model needs the reasoning of every earlier assistant message: send back the 'reasoning_content' returned with that message.Reproduction: multi-turn session on this model. The failure appears on a follow-up turn after an earlier assistant turn did substantial reasoning/generation (short trivial turns sometimes pass). Trace IDs available on request.
This matches the root cause analyzed in #877 (reasoning_content dropped when replaying prior assistant turns as history). The upstream only enforces the requirement once the prior turn produced real reasoning, which explains the intermittent nature.
Workaround in the meantime: switch to a model that does not return
reasoning_content(e.g. a non-thinking model) for multi-turn sessions.