Claude Fable 5.2 fast mode pricing: $16 in, $80 out

What fast mode costs

Fast mode on Fable 5.2 costs $16 per million input tokens and $80 per million output, exactly double the standard $8 and $40, per Anthropic's pricing page. In exchange, Anthropic promises up to 2.5 times higher output tokens per second. Mythos 5.2 fast mode costs the same but is only open to organizations in Anthropic's verification programs.

How fast mode stacks with discounts

• Prompt caching multipliers apply on top of fast mode prices.
• US-only inference (inference_geo: "us") adds 10% on top.
• Batch API does not work with fast mode.
• Fast mode pricing covers the full 1M context window, including requests over 200K input tokens.

To turn it on, set speed: "fast" on a Messages API request. No beta header is needed for Fable 5.2, per the release notes. Anthropic's launch post says fast mode is also available in Claude Code. It has not said in its launch materials how fast mode counts against Claude plan limits.

The odd overlap with Opus 5.5

Opus 5.5 fast mode costs $8 and $40, the same as standard Fable 5.2. So for the same bill you choose between a faster Opus 5.5 and a smarter but standard-speed Fable 5.2. Anthropic reports Fable 5.2 at High effort roughly matches Opus 5.5 at Max effort on FrontierCode for about half the cost per task, which favors Fable when quality matters more than latency.

Verdict

Skip Fable 5.2 fast mode for batch-style and background work: it doubles the bill, and the Batch API halves it instead. Use it for interactive agent loops where waiting costs more than tokens. If you need speed on a budget, Opus 5.5 fast mode at $8 and $40 is the better deal.

Full launch coverage: Claude Fable 5.2 and Mythos 5.2: prices, limits and what changed.