Simon Willison'sLLMA large language model — the neural network behind tools like Claude and ChatGPT, trained on huge amounts of text to predict what comes next.Full definition →-gemini 0.35 adds Gemini 4 Argon to LLM, a tool for calling language models. His release note says the model became available through the Gemini API on October 13.
The plugin connects through a Gemini API key and exposes low, medium and high thinking levels for Argon. Existing LLM users can therefore run the model from their command-line workflow and choose how much reasoning it performs before responding. The plugin documentation specifies those controls and leaves the model's default in place when no level is selected.
Key takeaways · AI-distilled
Simon Willison's llm-gemini 0.35 plugin adds Gemini 4 Argon, which is available through the Gemini API and to customers of Google's $100/month subscription plan.
At $2/$10 per million input/output tokenThe chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.Full definition →, Argon launches at GPT-6 Sol's price, half of Claude Opus 5.5 and a fifth of Claude Fable 5.1 and GPT-6 Astra; after the December 1 doubling it matches Opus 5.5.
Willison points to Artificial Analysis's total benchmarkA standard public test set for comparing AI models — the shared scoreboards behind every "model X beats model Y" claim.Full definition → run cost as the better comparison: $2,407 for Argon, close to GLM-5.3 and well below the latest OpenAI, Anthropic and Meta models.
Terms in this piece · Glossary
LLM — A large language model — the neural network behind tools like Claude and ChatGPT, trained on huge amounts of text to predict what comes next.
token — The chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.
benchmark — A standard public test set for comparing AI models — the shared scoreboards behind every "model X beats model Y" claim.
