LLM CLI 0.32 Released: Adds Reasoning Traces and Server-Side Tools
Simon Willison · rss · 2026-08-05
Simon Willison has released version 0.32 of the LLM CLI tool, a major upgrade that gives it powerful agent framework capabilities.
Key New Features:
- Visible Reasoning: Displays reasoning traces to standard error without polluting standard output, preserving pipe functionality.
- Server-Side Tools: Native support for OpenAI's Code Interpreter and WebSearch. The Anthropic plugin also supports WebSearch, CodeExecution, and MCP tools.
- Git-like Logging: Introduces a content-addressable SQLite message store to optimize multi-turn conversation logs and eliminate duplicate JSON payload overhead.
- Python API Upgrade: Allows passing complete message histories directly and introduces a new streaming events API (separating reasoning, text, and tool calls).
Additionally, the default model is now GPT-5.6 Luna, and the new llm openai endpoint command allows running one-off prompts against any OpenAI-compatible API.
More from coding & agent
- Anthropic Releases Free 1-Hour Workshop on Loop Engineering — goyalshaliniuk · 2026-08-05
- OpenAI's Codex Shows Monopolistic Trend, Squeezing Native Agent Frameworks — vista8 · 2026-08-05
- Microsoft's AgentStream: Evaluating Self-Evolving LLM Agents in Streaming Tasks — microsoft · 2026-08-05
- AI Agent Learns New Skills Autonomously: 8-Agent Loop Scours GitHub for Workflows — alexcovo_eth · 2026-08-05
- Pitfall: Claude API Skill Descriptions May Trigger Safety Classifiers — voooooogel · 2026-08-05
- Dev Workflow: 90% Codex, Drops Kimi K3 for GPT-5.6 Luna Max — DeryaTR_ · 2026-08-05