OpenClaw 2026.4.20 strengthens the default system prompt, adds tiered model pricing, and enforces session maintenance caps to keep your gateway healthy.

Highlights

📝 Smarter Default Prompts

The default system prompt and GPT-5 overlay have been strengthened with clearer completion bias, live-state checks, weak-result recovery, and verification-before-final guidance. Your agent will be more decisive and less likely to offer plans instead of results.

💰 Tiered Model Pricing

OpenClaw now supports tiered model pricing from cached catalogs and configured models. Bundled Moonshot Kimi K2.6/K2.5 cost estimates are included for accurate token-usage reports.

🧹 Session Maintenance

Built-in session entry cap and age pruning is now enforced by default. Oversized stores are pruned at load time so accumulated cron/executor session backlogs cannot OOM the gateway.

🔄 Cron State Separation

Cron runtime execution state is now split into jobs-state.json so jobs.json stays stable for git-tracked job definitions.

🌙 Kimi K2.6 Default

Moonshot setup, web search, and media understanding now default to kimi-k2.6 while keeping k2.5 for compatibility.

Key Fixes

YOLO exec no longer rejects gateway-host scripts in full security mode
Discord /think autocomplete only shows adaptive for providers that support it
Sessions reset now clears auto-sourced model overrides properly
Session cost compaction bug fixed (costs were being compounded multiple times)

How to Update

npm install -g openclaw@latest
openclaw gateway restart