# llm-anthropic 0.31 adds support for Anthropic Fable 5.2 and Mythos 5.2 models

> **Key takeaway:** The llm-anthropic plugin update enables CLI access to Anthropic's new Fable 5.2 and Mythos 5.2 models, which feature improved pricing and reasoning capabilities.

**Published:** 2026-11-09T23:58:46+00:00 
**Source:** Simon Willison Web & Open Source 
**Category:** software-engineering 
**Canonical URL:** https://fosswire.org/news/llm-anthropic-031-adds-support-for-anthropic-fable-52-and-my.html 

## Summary
The llm-anthropic Python plugin has been updated to version 0.31 to support the newly released Anthropic Fable 5.2 and Mythos 5.2 models. This update allows users of the llm command-line tool to interact with these models directly from the terminal using their Anthropic API keys.

Fable 5.2 introduces a tiered reasoning system with five distinct levels: low, medium, high, xhigh, and max. While the model offers improved performance and a 20% reduction in pricing compared to Fable 5.1, testing has identified a potential issue where the 'max' reasoning level can exhaust token limits during extended processing tasks.

## Technical details of the llm-anthropic 0.31 update
- **Model Integration:** The update adds support for the Claude-Fable-5.2 and Mythos 5.2 model identifiers to the llm CLI ecosystem.
- **Reasoning Tiers:** Users can specify five different reasoning levels (low, medium, high, xhigh, and max) when prompting the Fable 5.2 model.
- **Pricing Adjustments:** The new Fable 5.2 models are priced at $8 per million input tokens and $40 per million output tokens, down from the previous $10/$50 rate for Fable 5.1.
- **Observed Token Bug:** Testing indicates that the 'max' reasoning level may encounter a bug where it consumes up to 128,000 tokens during thought processes without returning a final response.

## Why it matters
This update provides software engineers and AI developers with streamlined CLI access to high-performance reasoning models via the llm tool. The reduced pricing for Fable 5.2 makes high-tier model integration more cost-effective for automated workflows and local development environments.