AivexaNewsSearch
AI news for builders and product teamsChecked every hour

llm-anthropic 0.31

Collected Nov 10, 2026

Simon Willison released version 0.31 of llm-anthropic, the plugin that connects his LLM command-line tool to Anthropic's Claude models, adding support for two new Anthropic models: Fable 5.2 and Mythos 5.2, both released the same day. Fable 5.2 currently sits first on the Artificial Analysis rankings, according to the release. Willison notes that Anthropic says the model "solved two open problems in the physics of black holes," which he reads wryly as evidence their internal benchmarks have grown more ambitious.

Using the new models takes three commands: install the LLM tool with uv, install the llm-anthropic plugin, store an Anthropic key with `llm keys set anthropic`, then prompt `llm -m claude-fable-5.2`. Willison's example prompt asks for an SVG of a pelican riding a bicycle, part of his long-running informal test of model drawing ability.

He ran that pelican prompt at all five reasoning effort levels — low, medium, high, xhigh and max — and published the results. The best output came from xhigh. The max level failed to return an image at all, exhausting its token budget after 128,000 tokens of thinking. Willison says this is the same failure he saw with Opus 5.5, but it is new for the Fable line: both Fable 5 and Fable 5.1 returned max-effort pelicans without trouble.

Pricing is $8 per million input tokens and $40 per million output tokens, a 20% cut from Fable 5.1 at $10/$50. Willison also used Fable 5.2 to help track down the source of a mysterious beep.

For developers, reasoning effort levels are increasingly a tuning knob as much as model choice: higher effort buys better answers but spends tokens on thinking before any output appears. A max setting that burns its full budget without producing the requested artifact suggests the effort-to-output balance can break at the top end. This is my own reading, not something the release states. The 20% price reduction is the more concrete change for teams budgeting API spend.

Why it matters: anyone piping Anthropic models through the LLM CLI can switch to the new versions immediately with an upgrade, and the lower token prices reduce cost for existing Fable workloads. Teams relying on maximum reasoning effort should test before moving to 5.2, since the max-level failure Willison reports may affect production prompts that expect an output rather than a truncated reasoning trace — an inference from his single test rather than a documented limitation.

Read at Simon Willison

Based on reporting from the original publisher. Visit the source for full context and later updates.

Publisher excerpt

Release: llm-anthropic 0.31 Anthropic released Fable 5.2 and Mythos 5.2 today. Fable 5.2 is currently ranked first on Artificial Analysis . Anthropic say it "solved two open problems in the physics of black holes", so clearly their internal benchmarks have become more ambitious! I've added the new models to llm-anthropic , so you can now run prompts through Fable 5.2 like this: uv tool install llm llm install llm-anthropic llm keys set anthropic # paste key here llm -m claude-fable-5.2 "Generate an SVG of a pelican