# Claude Fable 5.2 is out: faster, cheaper than 5.1

> Claude Fable 5.2 costs $8/$40 per million tokens, runs up to 30% faster than 5.1, and adds a fast mode. Here is what founders should test this week.

- URL: https://tools.launchllama.co/blog/claude-fable-5-2-faster-cheaper-than-5-1
- Published: 2026-11-10
- Publisher: Launch Llama (https://tools.launchllama.co)

TL;DR
- Claude Fable 5.2 and Mythos 5.2 are the same underlying model. Fable is open to all users with strong cyber and biology safeguards, and Mythos is for approved organizations.
- Price is $8 per million input tokens and $40 per million output tokens. Output is up to 30% faster than 5.1, and an optional fast mode is up to 2.5x faster at $16/$80.
- Anthropic still recommends starting with Claude Opus 5.5 for most coding and knowledge tasks, and moving up to Fable only for the hardest, most open-ended work.

## What happened

On November 9, 2026, Anthropic announced Claude Fable 5.2 and Claude Mythos 5.2. It says they improve on coding, reasoning, and vision, and communicate more clearly than earlier versions. They are the same underlying model. Fable has strong cybersecurity and biology safeguards, which lets Anthropic offer it to all users. Mythos is open to organizations that apply through the expanded Life Sciences and Cyber Verification Programs.

Pricing is $8 per million input tokens and $40 per million output tokens, which Anthropic describes as a lower price than Fable 5.1 and Mythos 5.1. Output is up to 30% faster than those versions. An optional fast mode generates tokens up to 2.5 times faster for $16 per million input tokens and $80 per million output tokens.

These are Anthropic's own benchmark results. On FrontierCode 2.0, Fable 5.2 scores 49.1% at Xhigh effort, which Anthropic calls the highest of any released model it tested. At High effort it scores 45.3% against Opus 5.5's best of 45.5%, for roughly half the cost per task. On BenchCAD 2.1 at Max effort it scores 83.5, against 65.3 for Opus 5.5 and 44.4 for Fable 5.1. Anthropic also says Medium effort, the default in Claude apps and Claude Code, uses fewer tokens than Fable 5.1 on hard tasks.

## What this means for founders

- **Cost per task may matter more than price per token.** Anthropic says Fable 5.2 at High effort matches Opus 5.5's best FrontierCode score for about half the cost per task. It also says Medium effort on WANDR matches Opus 5.5's top score at about 40% of the cost. Run your own tasks before trusting either number.
- **Don't switch everything.** Anthropic recommends starting with Opus 5.5 as the daily driver and going up to Fable or Mythos only for the most ambitious, open-ended technical work. Going down to Sonnet 5.5 or Haiku 5.5 is the cheaper path for simpler jobs.
- **Long unattended agent work is the pitch.** Anthropic describes an eight-day run with over 1,000 subagents that produced a working operating system, and an Uber engineer describes a 10 to 12 hour security audit with about 170 reviewer agents. These are vendor and customer accounts, not independent tests.
- **Fast mode costs double.** At $16/$80, it is twice the standard price for up to 2.5x speed. It is only worth it where waiting on the model costs you more than the extra tokens.
- **Security work hits a wall.** Fable 5.2's safeguards will block most cybersecurity-related tasks. If you build security or biotech products, you need to apply for Mythos access, and approval is not guaranteed by the source.

## What the Launch Llama directory shows

Live numbers from Launch Llama for listings that describe themselves as "ai coding assistant", "large language model", "ai model api", "ai agent" tools, as of 2026-11-10.

SignalLaunch Llama dataListings in this space320New in the last 30 days44With a free or freemium plan82%Most upvoted by foundersKiloClaw (774 upvotes)

Launch Llama lists 320 matching tools in this space, 44 of them live in the last 30 days, and 82% have a free or freemium plan, so the tooling around AI coding agents is crowded and still growing fast.

## How to try it today

- Pick one hard task you already run on Opus 5.5, such as a multi-file refactor, a migration, or a long agent job.
- Run it on Fable 5.2 at the default Medium effort in Claude apps or Claude Code. Note tokens used, cost, and whether the result needed fewer corrections.
- If the result falls short, rerun at High effort and compare cost per task against Opus 5.5, since that is where Anthropic's cost claims are made.
- Only test fast mode if latency is the bottleneck. It is priced at $16/$80 per million tokens.
- If you work in cybersecurity or biomedical research, apply for Mythos through the Life Sciences and Cyber Verification Programs.

Copy-paste starter prompt:

Review this codebase and propose a migration plan for [service name] from [old language] to [new language]. List risks, a step-by-step order of work, and the tests you would write first. Flag anything you are unsure about instead of guessing.

## Related tools on Launch Llama

Tools from the Launch Llama directory that fit with this news:

ToolWhat it doesWhy it matters hereClaude Fable 5.1For developers: Claude Fable 5.1 powers advanced coding and knowledge work so you ship faster and unlock early scientific AI capabilities.This is the previous version of the model, so you can compare your current Fable 5.1 workflow against 5.2.Agent IslandFor developers: Agent Island monitors Claude, Codex, Grok, and Cursor sessions locally so you track AI coding progress without telemetry.Fable 5.2 is pitched for long agent runs, and Agent Island tracks Claude and other coding agent sessions locally.Agent CalloutFor developers: Agent Callout adds voice alerts to AI coding agents so you monitor progress without leaving your terminal.Runs described for Fable 5.2 last hours to days, and Agent Callout adds voice alerts to coding agents so you know when they need you.

[Browse all tools](https://tools.launchllama.co/tools) or [list your own tool](https://tools.launchllama.co/submit).

## Llama's take

I would try Fable 5.2 now if you already pay for heavy agentic coding on Opus 5.5 and have one painful, long task to throw at it. The cost-per-task claims are the interesting part, and they are cheap to check on your own workload. Anthropic itself says to start with Opus 5.5 for most work, so I would not move everything over.

I would wait if your work is routine chat, small edits, or simple automation. Sonnet 5.5 and Haiku 5.5 are the stated lower tier, and Fable is not aimed at those jobs. Also treat every benchmark here as Anthropic's own reporting. The source points to a system card for methodology, and I have not seen independent results.

The one thing to watch is real cost per finished task on your own repos, not the per-token price. That decides whether Fable 5.2 saves you money or just feels better to use.

## FAQ

### How much does Claude Fable 5.2 cost?

Standard pricing is $8 per million input tokens and $40 per million output tokens, which Anthropic says is lower than Fable 5.1. Fast mode costs $16 per million input tokens and $80 per million output tokens.

### What is the difference between Claude Fable 5.2 and Mythos 5.2?

They are the same underlying model. Fable has strong cybersecurity and biology safeguards and is available to all users. Mythos is available to organizations that apply through Anthropic's Life Sciences and Cyber Verification Programs.

### Should I use Claude Fable 5.2 or Opus 5.5?

Anthropic recommends starting with Opus 5.5 for most coding and knowledge tasks. Move up to Fable or Mythos for the most ambitious, open-ended technical work, or down to Sonnet 5.5 or Haiku 5.5 for simpler jobs.

## Sources

- [Claude Fable 5.2 and Mythos 5.2](https://www.anthropic.com/claude-fable-and-mythos-5-2) (www.anthropic.com)
- [Hacker News discussion](https://news.ycombinator.com/item?id=50429906)

---

More for agents: https://tools.launchllama.co/llms.txt · https://tools.launchllama.co/catalog.md · https://tools.launchllama.co/mcp
