llm CLI releases 0.34 and 0.35: response-duration metrics and GPT-6 Astra model support

Simon Willison's open-source llm CLI shipped two releases: 0.34 (2026-09-02) adds response-duration metrics to log output, contributed bug fixes, and a significant performance improvement to llm logs, alongside llm-openrouter 0.7.1; 0.35 (2026-09-07) adds an…

Version 0.34 of the open-source llm command-line tool for interacting with large language models was released on 2026-09-02. Its headline change adds response duration in milliseconds and human-readable form to the llm logs --usage Markdown output, plus a new duration_ms field in llm logs --short (issue #1653). The release also includes several contributed bug fixes and a significant performance improvement to llm logs (faster log querying), and shipped alongside the related llm-openrouter 0.7.1 release. On 2026-09-07, llm 0.35 followed, adding an OpenAI model definition for GPT-6 Astra under the model ID gpt-6-astra; the 0.35 post provided no other release details.

• llm 0.34 was released on 2026-09-02 by Simon Willison's llm project, an open-source CLI for interacting with large language models.
• llm logs --usage Markdown output now shows response duration in milliseconds and human-readable form (llm 0.34).
• llm logs --short gains a new duration_ms field, tracked as issue #1653 (llm 0.34).
• llm 0.34 includes several contributed bug fixes and a significant performance improvement to llm logs, described as faster log querying.
• llm-openrouter 0.7.1 shipped alongside llm 0.34.
• llm 0.35 was released on 2026-09-07.
• llm 0.35 adds an OpenAI model definition for GPT-6 Astra under the model ID gpt-6-astra.
• The llm 0.35 post provided no other release details.