One setting boosted OpenClaw's score 44%: feed reasoning content back to the model
TheZachMueller · x · 2026-11-20
joelniklaus found that OpenClaw drops the model's earlier thinking when talking to custom OpenAI-compatible endpoints — no error, no warning — while reasoning models expect their prior thoughts in the conversation history. Flipping requiresReasoningContentOnAssistantMessages: true raised the return rate on GLM-5.3-Flash and DeepSeek-V4.1-Flash from 3-7% to nearly 100%. With its notes, OpenClaw needs about a third fewer model calls and hits time limits far less often; he's asked steipete and vincentkoc to make it the default.
More from coding & agent
- Perplexity: auto skill config cut tool-call failures by 11% in production test — perplexity_ai · 2026-11-20
- Perplexity: 72.9% of users gained personal skills with autonomous config vs 6.5% manual — perplexity_ai · 2026-11-20
- Perplexity's Brain reads full session records and abstains when evidence is thin — perplexity_ai · 2026-11-20
- Perplexity diagrams Brain's four-stage loop: evidence, reflection, verification, activation — perplexity_ai · 2026-11-20
- Perplexity's Brain now auto-creates self-improving agent skills from your past sessions — perplexity_ai · 2026-11-20
- Overmind launches autonomous AI lab: fine-tune your own model from production agent traces — jasonkneen · 2026-11-20