Pricing · Updated Nov 9, 2026
Claude Fable 5.2 pricing: $8 in, $40 out, and what a month really costs
Claude Fable 5.2 costs $8 per million input tokens and $40 per million output tokens on the Claude API. That is 20% below Fable 5.1's $10 / $50, and twice Opus 5.5's $4 / $20. Below: every rate from Anthropic's pricing page, three monthly bills worked out line by line, and the five settings that move the total most.
Input
$8 / MTok
Output
$40 / MTok
Cache read
$0.20 / MTok
Batch
50% off
Fable 5.2 API price table
From Anthropic's pricing page. Model ID claude-fable-5-2, 1M token context, 128K max output. Claude Mythos 5.2 has the same rates but limited access.
| Item | Price | Note |
|---|---|---|
| Input tokens | $8 / MTok | Was $10 on Fable 5.1 |
| Output tokens (thinking included) | $40 / MTok | Was $50 on Fable 5.1 |
| Cache write, 5 minutes | $10 / MTok | 1.25x input |
| Cache write, 1 hour | $16 / MTok | 2x input |
| Cache read (hit or refresh) | $0.20 / MTok | 2.5% of input. Most Claude models pay 10% |
| Batch API input / output | $4 / $20 per MTok | 50% off input and output |
| Fast mode input / output | $16 / $80 per MTok | Up to 2.5x faster generation. Not combinable with Batch |
| US-only inference (inference_geo: "us") | 1.1x every rate above | Global routing is the default and costs nothing extra |
| Prompts over 200K tokens | Same rates | The full 1M window bills at standard pricing |
Multipliers stack. A cache read in fast mode with US-only inference costs the fast-mode input rate × 2.5% × 1.1. The one exception: fast mode doesn't work with the Batch API.
Fable 5.2 against Opus, Sonnet and Haiku 5.5
Per million tokens, Claude API, global routing. All four current models have a 1M context window and 128K max output.
| Model | Input | Output | Cache read | Batch in / out |
|---|---|---|---|---|
Claude Fable 5.2claude-fable-5-2 | $8 | $40 | $0.20 | $4 / $20 |
Claude Fable 5.1claude-fable-5-1 | $10 | $50 | $0.25 | $5 / $25 |
Claude Opus 5.5claude-opus-5-5 | $4 | $20 | $0.20 | $2 / $10 |
Claude Sonnet 5.5claude-sonnet-5-5 | $2 | $10 | $0.10 | $1 / $5 |
Claude Haiku 5.5 (prompts up to 100K)claude-haiku-5-5 | $0.10 | $0.50 | $0.01 | $0.05 / $0.25 |
Claude Haiku 5.5 (prompts over 100K)claude-haiku-5-5 | $0.50 | $2.50 | $0.05 | $0.25 / $1.25 |
Fable 5.2 and Opus 5.5 charge the same $0.20 per million for cache reads, because Fable's cache discount is 2.5% of input and Opus's is 5%. On a workload that is mostly cached input, the gap between them shrinks toward the output price. For every provider, not only Anthropic, see our LLM API pricing comparison and the monthly AI model price index.
What Fable 5.2 costs per month: three worked examples
List prices only, no discounts or credits. The token counts are example sizes, not measurements. Swap in numbers from your own usage logs.
1. A coding agent, 2,000 tasks a month
Each task reads a 120K-token repo context from a warm cache, sends 30K fresh input tokens and writes 15K output tokens. Cache writes left out to keep the math readable.
- Fable 5.2 per task
- 0.12 × $0.20 + 0.03 × $8 + 0.015 × $40 = $0.864
- Fable 5.2 per month
- $1,728
- Opus 5.5 per month
- $888 ($0.444 per task)
- Sonnet 5.5 per month
- $444 ($0.222 per task)
Output is 69% of the Fable bill here. Effort setting moves output more than anything else, so set it on purpose.
2. A nightly document review on the Batch API
10,000 documents a month, 20K input tokens and 2K output tokens each, nothing cached.
- Fable 5.2, standard
- 10,000 × (0.02 × $8 + 0.002 × $40) = $2,400
- Fable 5.2, Batch API
- $1,200
- Opus 5.5, Batch API
- $600
If nobody is waiting for the answer, Batch halves the bill with no code change beyond the endpoint.
3. An interactive assistant on fast mode
500 sessions a month, 100K input tokens and 10K output tokens each, nothing cached.
- Fable 5.2, standard
- 500 × (0.1 × $8 + 0.01 × $40) = $600
- Fable 5.2, fast mode
- 500 × (0.1 × $16 + 0.01 × $80) = $1,200
- Fast mode with US-only inference
- $1,320
Fast mode doubles the bill. Pay it where a person sits watching the output, not in background jobs.
How to cut your Fable 5.2 bill
1. Route by task
The price ladder runs $0.10, $2, $4 and $8 per million input tokens across Haiku 5.5, Sonnet 5.5, Opus 5.5 and Fable 5.2. Anthropic's own advice is to start on Opus 5.5 and move up to Fable 5.2 only when evals at higher effort fall short. Send classification and extraction down to Haiku, keep Fable for the long agent runs that fail elsewhere.
2. Cache the long prefix
A cache read on Fable 5.2 costs $0.20 per million tokens, 2.5% of the $8 input rate. A 5-minute write costs 1.25x input, so it pays back on the first read. If every agent turn re-sends the same system prompt, tool list and repo map, caching them is the biggest single cut you can make. Read the prompt caching docs for breakpoints.
3. Batch anything that can wait
The Batch API takes 50% off input and output. Fable 5.2 on Batch is $4 / $20, which is Opus 5.5's standard price. Nightly evals, backfills and document reviews fit here. Fast mode doesn't work with Batch.
4. Set effort, don't inherit it
Fable 5.2 defaults to
higheffort on the API. The launch post saysmediumis the default in Claude apps and Claude Code. Thinking tokens bill as output at $40 per million, so an unset effort is a cost decision you didn't make. Anthropic reports that Fable 5.2 at high effort matched Opus 5.5's best FrontierCode 2.0 score for roughly half the cost per task, so tryhighbeforexhigh. See the effort docs.5. Skip US-only inference unless you need it
Setting
inference_geo: "us"adds 10% to every token category. Use it when a contract requires US data processing, not by default.
Deciding between the two top models? The Fable 5.2 vs Opus 5.5 comparison works out cost per successful task, and the migration guide covers the switch.
Launch offers for Claude subscribers
This is for Claude plans, not the API. Per Anthropic's help center article:
- Pro and standard Team seats can spend up to 50% of their weekly usage limits on Fable models from November 9 to November 16, 2026 (11:59 pm PT), and get one Fable reset that expires the same day. After that, Fable usage moves to usage credits if you have them turned on.
- Max, Ultra, premium Team seats and premium seats on legacy seat-based Enterprise plans get resets by tenure: one under six months, two for six to 12 months, three past 12 months. These expire March 8, 2027 (11:59 pm PT).
- A reset restores Fable usage. It doesn't raise your limits. The Free plan, HIPAA-ready Enterprise plans, and standard, self-serve or usage-based Enterprise seats don't get resets.
- Paid billing had to be in place before November 9, 2026 at 10:00 am PT.
Want the bill under control from day one?
Routing, caching and effort settings are engineering work that someone has to own. AY Automate places AI engineers inside your team, from $3K a month with a 90-day replacement guarantee.
FAQ
How much does Claude Fable 5.2 cost?
$8 per million input tokens and $40 per million output tokens on the Claude API. Cache reads are $0.20 per million, the Batch API is half price at $4 / $20, and fast mode is $16 / $80.
Is Fable 5.2 cheaper than Fable 5.1?
Yes, by 20%. Fable 5.1 was $10 input and $50 output per million tokens. Fable 5.2 is $8 and $40. Cache reads dropped from $0.25 to $0.20 per million.
How does Fable 5.2 pricing compare with Opus 5.5?
Fable 5.2 costs twice as much per token: $8 / $40 against $4 / $20. Cache reads cost the same $0.20 per million on both. Anthropic reports that on FrontierCode 2.0, Fable 5.2 at high effort matched Opus 5.5's best score for roughly half the cost per task, so cost per finished task can differ from the token ratio.
What does Fable 5.2 fast mode cost?
$16 per million input tokens and $80 per million output tokens, double the standard rate, for up to 2.5x faster generation. It is available on the Claude API, Claude Models on Google Cloud and Microsoft Foundry for Anthropic-hosted deployments. On the Google-managed option, Google sets the price. It can't be combined with the Batch API.
Is there a long-context surcharge on Fable 5.2?
No. Anthropic bills the full 1M token context window at standard rates. A 900K-token request pays the same per-token price as a 9K-token request.
Can I use Claude Fable 5.2 for free?
Not on the Free plan. Launch offers go to paid plans: Pro and standard Team seats can use up to 50% of their weekly limits on Fable models until November 16, 2026, plus one reset. Max, Ultra and premium seats get one to three resets depending on how long they have subscribed. Billing had to be in place before November 9, 2026 at 10:00 am PT.
What does Fable 5.2 cost on Amazon Bedrock or Google Cloud?
Anthropic's pricing page says partner-operated platforms, Amazon Bedrock and Claude Models on Google Cloud (Google managed), set their own regional prices. Check the AWS and Google Cloud pricing pages. The Claude API rates on this page apply to the Claude API, Claude Platform on AWS and Claude Models on Google Cloud (Anthropic managed).
Keep reading
- Claude Fable 5.2 vs Opus 5.5
- Migrate Claude Opus 5.5 to Fable 5.2
- Claude Fable 5.2 for engineering teams
- Hire Claude Fable 5.2 engineers
- Claude Fable 5.1 pricing and benchmarks