Claude Fable 5.2 costs $8 per million input tokens and $40 per million output tokens, 20% less than Fable 5.1 and exactly 2x Opus 5.5. Here is the full rate card with fast mode, cache and batch prices, Anthropic's cost-per-task data, worked examples with the math shown, and what AI search visibility checks on Claude cost.
Client result: Wearview went from $0 to $36K MRR in 8 months.
This post is part of our AI SEO guide. Related reading: ChatGPT Product Feed Requirements, AI Licensing Deals Boost ChatGPT Citations 48%, Only, Google AI Mode Ads Do Not Buy You a.
Adel tracks AI citation rates across ChatGPT, Perplexity, Gemini, and AI Overviews. He turns raw visibility data into actionable insights that guide our optimization strategy.
Full Bio →Claude Fable 5.2 pricing is $8 per million input tokens and $40 per million output tokens on the Claude API. That is 20% below Fable 5.1, which cost $10 and $50. Fast mode costs $16 and $80, the Batch API halves the standard rate, and cache reads cost $0.20 per million tokens.
When we checked Google's top results for this query at 21:29 UTC on launch day, several of them still showed Fable 5 or Fable 5.1 prices. This page has the November 9 rate card, the comparison with Opus 5.5, Anthropic's own cost-per-task data, and worked examples with the math shown so you can check it.
The last section covers a cost marketers ask us about: what it takes to run AI search visibility checks on Claude with web search switched on.
Charts re-checked 2026-11-09 at 23:54 UTC. Facts checked 2026-11-09, between 21:00 and 21:30 UTC, against Anthropic's pricing page and announcement.
Every number in this table comes from Anthropic's pricing page. Prices are in USD per million tokens (MTok).
Five billing rules apply on top of the table:
inference_geo: "us" applies a 1.1x multiplier to every token category, cache reads included.Need the specs, release date and benchmarks instead? See our Claude Fable 5.2 fact sheet.
Claude Fable 5.2 costs exactly twice as much as Opus 5.5 per token, on both input and output. It costs 20% less than Fable 5.1. All rows are from the pricing page.
Look at the cache read column. Fable 5.2 and Opus 5.5 both charge $0.20 per million cached tokens. If most of your input is a reused system prompt or document, the input-side gap between the two models gets much smaller than 2x. Output stays at 2x.
Claude Mythos 5.2 is listed at the same prices as Fable 5.2. Direct access is limited to verified organizations.
Fable 5.2 also costs less per token than GPT-6 Astra, which OpenAI lists at $10 and $50, and twice as much as Opus 5.5.
Fast mode for Claude Fable 5.2 costs $16 per million input tokens and $80 per million output tokens, double the standard rate. Anthropic says it delivers "up to 2.5x higher output tokens per second," per the fast mode docs.
What you should know before you turn it on:
Fast mode is available on Claude Code, the Claude Platform, Claude Models on Google Cloud and Anthropic-hosted deployments on Microsoft Foundry, per the announcement. On Google Cloud's Google-managed option, Google sets its own fast mode prices.
Pay for fast mode when a person is waiting on the answer. For overnight agent runs, batch jobs or scheduled reports, standard speed or the Batch API costs a half or a quarter as much.
Not always. Per token, Fable 5.2 is 2x Opus 5.5. Per finished task, Anthropic's own tests show it can be cheaper, because it can reach the same score at a lower effort level. These numbers all come from the announcement, so treat them as vendor results.
Every comparison sets Fable 5.2 at a lower effort level against Opus 5.5 at its best. The saving comes from the lower effort setting. At max effort the picture changes.
Artificial Analysis puts Fable 5.2 at max effort first on its Intelligence Index with a score of 65, and also reports it costs $26.54 per index task at that setting. It calls the max-effort setup "particularly expensive" next to models of similar price. Checked 2026-11-09 at 21:28 UTC, and the same on 2026-11-09 at 23:54 UTC.
Effort changes the bill more than the rate card does. On Artificial Analysis, Fable 5.2 at high effort scores 62 for $9.08 per task. Opus 5.5 at max effort scores 59 for $10.04.
usage.output_tokens_details.thinking_tokens in each response to see how much of your bill is thinking.
Here are two examples you can check line by line. The token counts are assumptions we picked to keep the math readable. The prices are Anthropic's.
| Option | Input math | Output math | Total |
|---|---|---|---|
| Fable 5.2 standard | 200,000 × $8 / 1M = $1.60 | 10,000 × $40 / 1M = $0.40 | $2.00 |
| Fable 5.1 standard | 200,000 × $10 / 1M = $2.00 | 10,000 × $50 / 1M = $0.50 | $2.50 |
| Opus 5.5 standard | 200,000 × $4 / 1M = $0.80 | 10,000 × $20 / 1M = $0.20 | $1.00 |
| Fable 5.2 Batch API | 200,000 × $4 / 1M = $0.80 | 10,000 × $20 / 1M = $0.20 | $1.00 |
| Fable 5.2 fast mode | 200,000 × $16 / 1M = $3.20 | 10,000 × $80 / 1M = $0.80 | $4.00 |
The 10,000 output tokens include thinking. At higher effort, expect that number to grow.
Say an agent makes 100 requests. Each one reuses the same 150,000-token prefix of instructions and documents, adds 5,000 new input tokens and writes 2,000 output tokens. All requests land inside the 5-minute cache window, and each read refreshes it.
| Line | Math | Cost |
|---|---|---|
| First cache write, 5 minutes | 150,000 × $10 / 1M | $1.50 |
| 99 cache reads | 99 × 150,000 = 14.85M × $0.20 / 1M | $2.97 |
| New input | 100 × 5,000 = 500,000 × $8 / 1M | $4.00 |
| Output | 100 × 2,000 = 200,000 × $40 / 1M | $8.00 |
| Total with caching | $16.47 | |
| Same job without caching | 15.5M input × $8 / 1M + $8.00 output | $132.00 |
Caching cuts this job from $132.00 to $16.47, about 88% less. The same cached job on Opus 5.5 comes to $9.72: $0.75 write, $2.97 reads, $2.00 new input and $4.00 output. The cache reads cost the same on both models, so most of the gap comes from output tokens.
An AI search visibility check means asking Claude a question your buyers ask, with web search switched on, and recording whether your brand gets mentioned or cited. On the Claude API you pay twice for each check: once for tokens and once for the searches.
Anthropic's pricing page sets web search at "$10 USD per 1,000 searches, plus standard token costs for search-generated content." Three billing details matter here:
Assume you track 50 buyer prompts and run each one 3 times, because answers can differ between runs. That is 150 checks. Assume each check uses 3 searches, 30,000 input tokens including the search results, and 2,000 output tokens including thinking. These are our assumptions. Measure your own with the usage field, which reports web_search_requests next to token counts.
The input line is 30,000 × $8 / 1M = $0.24, the output line is 2,000 × $40 / 1M = $0.08, and 3 searches at $10 per 1,000 come to $0.03.
Two things follow from the table. Fable 5.2 through the Batch API costs the same as Opus 5.5 at standard rates, so at the same token counts a weekly tracking job can run on the stronger model for no extra money if it can wait. Fable 5.2's higher default effort can mean more output tokens, so measure before you switch. And if you want to know what Claude Fable 5.2 itself cites, you have to run the checks on Fable 5.2. Results from another model tell you about that model.
Dynamic filtering helps too. With the web_search_20260209 tool version and later, Claude filters results with code before they reach the context window, which Anthropic says "reduces token use on search-heavy requests." Code execution used this way carries no extra charge beyond tokens.
The API tool is not necessarily what a person sees in the Claude app. Anthropic documents the API tool. We found no page that says the consumer apps search the same way. Our guide on how to get cited by Claude covers what the docs say about citations, and our GEO service runs this tracking for clients across Claude, ChatGPT, Perplexity and Gemini.
A 150-check Claude visibility run costs about $52.50 on Fable 5.2 at standard rates under our assumptions, or $28.50 through the Batch API. Search fees are a small share of the bill. Input tokens are most of it.
On Max, Ultra and premium seats, Fable 5.2 draws on your plan's usage limits. On Pro plans and standard Team seats it normally runs on usage credits. For launch week, Anthropic's launch offer article says Pro plans and standard Team seats get all Fable models included from November 9 to November 16, 2026, at 11:59 pm PT, up to 50% of weekly usage limits. Those seats also get one Fable limit reset that expires at the same time. After that, Fable usage on those plans goes to usage credits if you have them turned on.
Max, Ultra, premium Team seats and premium seats on legacy seat-based Enterprise plans get one to three Fable limit resets, based on how long they have subscribed. Resets only go to accounts with paid billing in place before November 9, 2026, at 10:00 am PT. The launch offers don't cover the Free plan. Anthropic also warns that "Fable models use your usage limits faster than other Claude models."
Claude Fable 5.2 costs $8 per million input tokens and $40 per million output tokens on the Claude API. Cache reads cost $0.20 per million tokens and the Batch API charges $4 and $20. Prices come from Anthropic's pricing page, checked November 9, 2026.
Claude Fable 5.2 fast mode costs $16 per million input tokens and $80 per million output tokens. Anthropic says it gives up to 2.5x higher output tokens per second. You can't combine it with the Batch API, but prompt caching discounts still apply on top.
Yes, Claude Fable 5.2 is 20% cheaper than Fable 5.1 on both input and output. Fable 5.1 cost $10 and $50 per million tokens, against $8 and $40 for Fable 5.2. Cache reads dropped from $0.25 to $0.20.
Claude Fable 5.2 is priced at exactly 2x Opus 5.5 per token: $8 and $40 against $4 and $20. Anthropic doesn't explain the gap. It positions Fable 5.2 for demanding reasoning and long-horizon agent work, and recommends Opus 5.5 as the starting point for most workloads. Per task, Anthropic's tests show Fable 5.2 at lower effort can cost less than Opus 5.5 at its best.
Yes. The Batch API charges $4 per million input tokens and $20 per million output tokens for Claude Fable 5.2, half the standard rate. Fast mode is not available through the Batch API.
Web search costs $10 per 1,000 searches on the Claude API, plus token costs, and search results count as input tokens. A search that errors is not billed. Under our assumptions, one visibility check with 3 searches costs about $0.35 on Fable 5.2. Our GEO optimization service runs these checks across AI assistants for you.
Not as part of the launch offers, which don't cover the Free plan. Pro plans and standard Team seats get Fable models included until November 16, 2026, capped at 50% of weekly usage limits. On the API you pay per token, after the small free credit Anthropic gives new users to test the API.
Facts checked 2026-11-09. Sources: Claude Platform pricing, Anthropic, Introducing Claude Fable 5.2 and Claude Mythos 5.2, Claude Platform models overview, Fast mode docs, Adaptive thinking docs, Web search tool docs, Claude Fable 5.2 launch offers, Artificial Analysis, Claude Fable 5.2
| Price item | Claude Fable 5.2 |
|---|---|
| Input | $8 / MTok |
| Output | $40 / MTok |
| 5-minute cache write | $10 / MTok (1.25x input) |
| 1-hour cache write | $16 / MTok (2x input) |
| Cache read (hit or refresh) | $0.20 / MTok |
| Batch API input | $4 / MTok |
| Batch API output | $20 / MTok |
| Fast mode input | $16 / MTok |
| Fast mode output | $80 / MTok |
| Model | Input | Output | Cache read | Batch in / out |
|---|---|---|---|---|
| Claude Fable 5.2 | $8 | $40 | $0.20 | $4 / $20 |
| Claude Fable 5.1 | $10 | $50 | $0.25 | $5 / $25 |
| Claude Opus 5.5 | $4 | $20 | $0.20 | $2 / $10 |
| Claude Sonnet 5.5 | $2 | $10 | $0.10 | $1 / $5 |
| Claude Haiku 5.5 | from $0.10 | from $0.50 | from $0.01 | from $0.05 / $0.25 |
| Fast mode fact | Detail |
|---|---|
| How to use it | Set speed: "fast" on the request |
| What gets faster | Output tokens per second, not time to first token |
| Batch API | Not available with fast mode |
| Prompt caching | Cache multipliers apply on top of fast mode prices |
| Microsoft Foundry | A separate model ID, claude-fable-5-2-fast |
| Opus 5.5 fast mode | $8 / $40, the same as Fable 5.2 at standard speed |
| Benchmark | Anthropic's cost claim |
|---|---|
| FrontierCode 2.0 | At high effort, Fable 5.2 scores 45.3% vs Opus 5.5's best 45.5% "for roughly half the cost per task" |
| WANDR (agentic search) | At medium effort, Fable 5.2 "matches Opus 5.5's top score for about 40% of the cost per task" |
| Humanity's Last Exam | At low effort, Fable 5.2 "matches Opus 5.5's top score for less than one-tenth of the cost per task" |
| Chartography | At medium effort, Fable 5.2 scores 66.4% vs Opus 5.5 at max effort (64.4%) "for less than one-tenth of the cost per task" |
| Setup | Per-check math | Per check | 150 checks |
|---|---|---|---|
| Fable 5.2 standard | $0.24 input + $0.08 output + $0.03 search | $0.35 | $52.50 |
| Fable 5.2 Batch API | $0.12 input + $0.04 output + $0.03 search | $0.19 | $28.50 |
| Opus 5.5 standard | $0.12 input + $0.04 output + $0.03 search | $0.19 | $28.50 |