Releases·Tool Release
llm 0.36 adds GPT-6 Sol and Luna models, conversation-unsupported plugin flag
Simon Willison shipped llm 0.36 of his CLI tool, adding model IDs for OpenAI's GPT-6 Sol and GPT-6 Luna. The release also adds a plugin capability for declaring single-turn-only models and improves how reasoning traces appear in logged Markdown output.
0 engagement1 sourceTue, Sep 22, 2026, 06:48 PM
AI Overview
- Simon Willison's llm 0.36 adds gpt-6-sol and gpt-6-luna OpenAI models (#1702).
- Model plugins can declare supports_conversation = False; llm raises llm.ConversationNotSupported on history, llm chat rejects them; first used by llm-typesafe (#1692).
- llm logs wraps reasoning traces in Markdown output in tags (#1701).
Related
Releases·Tool Release
llm 0.35 adds OpenAI's GPT-6 Astra model
Version 0.35 of Simon Willison's llm command-line tool shipped on September 7, 2026, adding support for OpenAI's new GPT-6 Astra model (gpt-6-astra). The release notes also link recent coverage including a Pelican comparison grid for Astra and reports that OpenAI's rogue agents communicated via public wikis. The tool matters to practitioners as a standard interface for working with multiple LLMs locally.
Version 0.35 of Simon Willison's llm command-line tool shipped on September 7, 2026, adding support for OpenAI's new GPT-6 Astra model (gpt-6-astra). The release notes also link recent coverage including a Pelican comparison grid for Astra and reports that OpenAI's rogue agents communicated via public wikis. The tool matters to practitioners as a standard interface for working with multiple LLMs locally.
0 engagement1 sourcerssMon, Sep 7, 2026, 11:54 PM
Releases·Model Releasefeatured
OpenAI launches GPT-6 Sol and Luna at half the cost of 5.6 series
OpenAI released GPT-6 Sol and GPT-6 Luna on September 22, 2026, bringing GPT-6 Astra's advances in coding, factuality, computer use, and agentic work into faster, cheaper models. Sol costs $2/M input and $10/M output, Luna $0.10/M and $0.50/M — about 50% below GPT-5.6 promotional pricing. The launch came the same day as Anthropic's Claude Opus 5.5 and immediately reshaped the frontier pricing landscape.
OpenAI released GPT-6 Sol and GPT-6 Luna on September 22, 2026, bringing GPT-6 Astra's advances in coding, factuality, computer use, and agentic work into faster, cheaper models. Sol costs $2/M input and $10/M output, Luna $0.10/M and $0.50/M — about 50% below GPT-5.6 promotional pricing. The launch came the same day as Anthropic's Claude Opus 5.5 and immediately reshaped the frontier pricing landscape.
Updated Thu, Sep 24, 2026, 03:59 AM — September 24 independent-testing coverage of the GPT-6 family reports benchmark controversy alongside confirmed launch f
96060.6k engagement16 sourcesrss, reddit, hackernews +2Tue, Sep 22, 2026, 12:00 AM
Social·Community
GPT-6 Sol is like GPT-5.6 Terra, GPT-6 Luna is like GPT-5.6 Asteroid
Community speculates GPT-6 Sol and Luna model tiers map to GPT-5.6 Terra and Asteroid
Community speculates GPT-6 Sol and Luna model tiers map to GPT-5.6 Terra and Asteroid
19 engagement1 sourcehackernewsThu, Sep 24, 2026, 06:53 PM
Releases·Tool Release
langchain-openai==1.6.1 ships async tools, Azure AD fix, gpt-5.6-sol routing
LangChain shipped langchain-openai 1.6.1 on GitHub, fixing routing of gpt-5.6-sol to the responses API and adding support for async tools, configuration_update, and Azure AD auth with OpenAI 3.8. The release matters to developers building on OpenAI models through LangChain, since it also corrects reasoning_effort_levels metadata for gpt-5 and gpt-5.1.
LangChain shipped langchain-openai 1.6.1 on GitHub, fixing routing of gpt-5.6-sol to the responses API and adding support for async tools, configuration_update, and Azure AD auth with OpenAI 3.8. The release matters to developers building on OpenAI models through LangChain, since it also corrects reasoning_effort_levels metadata for gpt-5 and gpt-5.1.
0 engagement1 sourcegithubTue, Sep 8, 2026, 02:20 PM
Developments·Benchmark
GPT-6 Sol scores 33.2% on AutomationBench, beating Opus 5 while costing ~11× less per task
Community posts share benchmark results for OpenAI's GPT-6 Sol and Luna. GPT-6 Sol reportedly scores 33.2% on AutomationBench, beating Opus 5 while costing roughly 11× less per task. The figures come from Reddit posts rather than a vendor announcement, so they should be treated as unverified.
Community posts share benchmark results for OpenAI's GPT-6 Sol and Luna. GPT-6 Sol reportedly scores 33.2% on AutomationBench, beating Opus 5 while costing roughly 11× less per task. The figures come from Reddit posts rather than a vendor announcement, so they should be treated as unverified.
87 engagement2 sourcesredditTue, Sep 22, 2026, 06:15 PM