Claude Fable API Prices, Fable 5 to 5.2
Claude Fable 5.2 costs $8 per million input tokens and $40 per million output, down from $10 and $50. Fable 5.1 had already cut cache reads by 75%.

Anthropic released Claude Fable 5.2 on Nov 9, 2026, at $8 per million input tokens and $40 per million output tokens. The two Fable models before it cost $10 and $50. Fable 5.2 is the first Fable model to cut those prices, and Fable 5.1 was the first to cut a price of any kind, when it lowered the price of cache reads in September.
Fable 5 and 5.1 held $10 and $50
Claude Fable 5 launched on Jun 9 at $10 per million input tokens and $50 per million output tokens. Anthropic's pricing page lists its cache reads at $1 per million tokens, a tenth of its input price.
Claude Fable 5.1 followed on Sep 1 at the same input and output prices. Anthropic said in its launch post that "cache reads now cost 75% less, or $0.25 per million tokens". The post describes a cache read as the step "where the model reads inputs that have already been processed and stored".
Fable 5.2 cuts input, output and cache prices by 20%
Fable 5.2 lowers each of the three prices by a fifth. Its input costs $8, its output $40 and its cache reads $0.20, according to the launch post and the pricing page. Anthropic also added a fast mode for Fable 5.2 that generates tokens up to 2.5 times faster, at $16 per million input tokens and $80 per million output tokens.
Fable 5.2's cache reads now cost the same as Opus 5.5's
Claude Opus 5.5, released on Sep 22, costs $4 per million input tokens and $20 per million output tokens, half of Fable 5.2's prices. Its cache reads cost $0.20, the same as Fable 5.2's. Anthropic's pricing page sets a cache read at 2.5% of the input price on Fable 5.2 and 5.1, and at 5% on Opus 5.5. Fast mode on Opus 5.5 costs $8 and $40.
GPT-6 Astra still costs what Fable 5 did
OpenAI's GPT-6 Astra, released on Sep 3, lists at $10 per million input tokens and $50 per million output tokens for prompts up to 272,000 tokens, the same as Fable 5 and 5.1. Its cached input costs $1 per million tokens, five times Fable 5.2's cache-read price.
| Model | Released | Input, per 1M | Cache read, per 1M | Output, per 1M |
|---|---|---|---|---|
| Claude Fable 5 | Jun 9 | $10 | $1 | $50 |
| Claude Fable 5.1 | Sep 1 | $10 | $0.25 | $50 |
| Claude Fable 5.2 | Nov 9 | $8 | $0.20 | $40 |
| Claude Opus 5.5 | Sep 22 | $4 | $0.20 | $20 |
| GPT-6 Astra | Sep 3 | $10 | $1 | $50 |
Input price, $ per 1M tokens
Cache read price, $ per 1M tokens
Output price, $ per 1M tokens
Fast mode input price, $ per 1M tokens
Fast mode output price, $ per 1M tokens
Sources
- Anthropic, blog post, Jun 9, 2026Claude Fable 5 and Claude Mythos 5
- Anthropic, blog post, Sep 1, 2026Introducing Claude Fable 5.1 and Claude Mythos 5.1
- Anthropic, blog post, Nov 9, 2026Introducing Claude Fable 5.2 and Claude Mythos 5.2
- Anthropic, blog post, Sep 22, 2026Introducing Claude Opus 5.5
- Anthropic, pricing pageplatform.claude.com/docs/en/about-claude/pricing
- OpenAI, pricing pagedevelopers.openai.com/api/docs/pricing