Claude Fable 5.2, released on November 9, 2026, cuts the Fable price to $8 / $40 per million tokens, with $0.20 cache reads, $4 / $20 batch rates and a new $16 / $80 fast mode. The rate card, worked examples and Opus comparison now lead with Fable 5.2 and keep Fable 5.1 and Fable 5 for reference. The plan section adds the Fable 5.2 launch offers, Enterprise admin enablement and the Claude Code version for 5.2.
Claude Fable 5.2, the current Fable model, costs $8 per million input tokens and $40 per million output tokens on the API, 20% less than Fable 5.1 and Fable 5 ($10 / $50) and exactly 2× Claude Opus 5.5 ($4 / $20). Cache reads cost $0.20/MTok on Fable 5.2 (97.5% off, the same as Opus 5.5) versus $0.25 on Fable 5.1 and $1 on Fable 5, and the Batch API halves the base rates to $4 / $20. On Claude plans, Fable is included on Max and Ultra at up to 50% of the weekly limit; on Pro it runs on usage credits billed at API rates, after a Fable 5.2 launch week that ends November 16, 2026.
How much does Claude Fable 5 cost per token in 2026?
Claude Fable 5.2, the current Fable model, costs $8 per million input tokens (MTok) and $40 per million output tokens on the standard API, according to Anthropic’s official pricing documentation as of November 12, 2026; Fable 5.1 and Fable 5 stay at $10 / $50. That is the base rate before any caching or batch discount, and it applies across the full 1M-token context window with no long-context surcharge. Output is metered at 5× the input rate, the usual Claude shape at a higher absolute level than the rest of the lineup.
Fable 5.2, released on November 9, 2026, replaced Fable 5.1 (September 1, 2026) as the current Fable model; Fable 5.1 and Fable 5 remain available as older models. Fable 5.2 is 20% cheaper on every base row, and the cache read, which matters more than it sounds for agent workloads, falls to $0.20. For benchmarks and migration notes, see our Claude Fable 5.2 guide. The plan-side rules have also changed since this guide first appeared: Fable is now a standard part of Max and Ultra, with Pro on usage credits. This page is strictly about price and limits. For the June 2026 access interruption, see our write-up on the Fable 5 and Mythos 5 availability change; for where Fable sits among Anthropic’s models, see the Claude models map for 2026.
Prices on platform.claude.com can change without notice. Every figure here reflects the rate card as of November 12, 2026. Re-check the canonical pricing page before committing a large spend forecast.
What is the full Claude Fable 5 API rate card?
The full Fable rate card has more rows than most pricing summaries show. Competitors quote the headline $8/$40 and stop; the rows that actually decide your bill are the cache writes, the cache hit and the batch discount. Here is the canonical table for the current Fable 5.2, with Fable 5.1 and Fable 5 kept for reference, sourced to Anthropic’s platform pricing docs.
| Operation | Fable 5.2 (per MTok) | Fable 5.1 (per MTok) | Fable 5 (per MTok) | Multiplier vs base input | When it applies |
|---|---|---|---|---|---|
| Base input | $8.00 | $10.00 | $10.00 | 1× | Every uncached input token |
| Base output | $40.00 | $50.00 | $50.00 | 5× | Every generated token |
| Cache write (5-minute TTL) | $10.00 | $12.50 | $12.50 | 1.25× | First time you cache a prefix, short-lived |
| Cache write (1-hour TTL) | $16.00 | $20.00 | $20.00 | 2× | First time you cache a prefix, long-lived |
| Cache hit / read | $0.20 | $0.25 | $1.00 | 0.025× on 5.2 and 5.1; 0.1× on 5 | Reusing a cached prefix |
| Batch API (input) | $4.00 | $5.00 | $5.00 | 0.5× | Asynchronous batch jobs |
| Batch API (output) | $20.00 | $25.00 | $25.00 | 0.5× of output | Asynchronous batch jobs |
| Fast mode (input / output) | $16.00 / $80.00 | Not offered | Not offered | 2× standard | Opt-in, up to 2.5× output speed; not available with the Batch API |
Note that Mythos 5.2, Mythos 5.1 and Mythos 5 share the rate cards of their Fable counterparts, but access is limited: Anthropic offers Mythos 5.2 only to organizations verified through its verification programs, the Cyber Verification Program and the Life Sciences Verification Program. Do not assume you can simply call Mythos at these rates; the rate card matches, the access does not.
How do prompt caching and batch discounts change the real bill?
Prompt caching and batch processing are the two levers that move a Fable bill the most. On Fable 5.2 and 5.1, caching cuts repeated-context input cost by 97.5% (90% on Fable 5), and batch halves everything for work you don’t need in real time. The trick is knowing when each one pays for itself, because cache writes cost more than a plain input token up front.
When does prompt caching pay off on Fable 5?
Caching pays off after a single read with the 5-minute cache, or after two reads with the 1-hour cache, on every Fable version. The math is direct: a 5-minute cache write costs 1.25× base input ($10/MTok on Fable 5.2, $12.50 on 5.1 and 5), and every subsequent hit costs $0.20/MTok on Fable 5.2, $0.25/MTok on Fable 5.1 or $1/MTok on Fable 5. So caching a 1M-token prefix once and reading it once costs $10.20 on Fable 5.2, versus $16.00 for two uncached passes ($12.75 on Fable 5.1 and $13.50 on Fable 5, versus $20.00), already cheaper on the first reuse. The 1-hour cache writes at 2× ($16/MTok on Fable 5.2, $20 on 5.1 and 5), so it breaks even on the second read and wins clearly from the third onward. For long system prompts, large document context, or agent loops that re-send the same tool definitions every turn, caching is close to mandatory on an $8-input model, and Fable 5.2’s $0.20 reads make it more valuable still.
How much does the Batch API save?
The Batch API applies a flat 50% discount in both directions, taking Fable 5.2 to $4/MTok input and $20/MTok output ($5 / $25 on Fable 5.1 and Fable 5). Batch is for asynchronous work (evals, bulk classification, offline summarization, content pipelines) where you can tolerate the queued turnaround. If your workload is offline and high-volume, batch is often the single biggest lever, bigger than caching, because it discounts output too, and output is where Fable’s cost concentrates at 5× input.
5-minute cache: net positive after 1 reuse. 1-hour cache: net positive after 2 reuses. Below that, the write premium (1.25× / 2×) costs you. For one-shot calls with no shared prefix, skip caching entirely.
What does a real Claude Fable 5 bill look like?
A concrete job, 1M input tokens plus 200k output tokens, costs $16.00 uncached on Fable 5.2 ($20.00 on 5.1 and 5), drops to about $8.98 on Fable 5.2 ($11.23 on Fable 5.1, $11.90 on Fable 5) if most of the input is a cached prefix you reuse, and falls to $8.00 on the Batch API ($10.00 on 5.1 and 5). The example assumes a single request shape; your real distribution of cached vs fresh input will differ.
| Mode | Input cost | Output cost | Total |
|---|---|---|---|
| Fable 5.2: uncached, real-time | 1M × $8 = $8.00 | 200k × $40 = $8.00 | $16.00 |
| Fable 5.2: 900k cached read + 100k fresh | (0.9M × $0.20) + (0.1M × $8) = $0.98 | 200k × $40 = $8.00 | $8.98 |
| Fable 5.2: Batch API (uncached) | 1M × $4 = $4.00 | 200k × $20 = $4.00 | $8.00 |
| Uncached, real-time (Fable 5 or 5.1) | 1M × $10 = $10.00 | 200k × $50 = $10.00 | $20.00 |
| Fable 5.1: 900k cached read + 100k fresh | (0.9M × $0.25) + (0.1M × $10) = $1.23 | 200k × $50 = $10.00 | $11.23 |
| Fable 5: 900k cached read + 100k fresh | (0.9M × $1) + (0.1M × $10) = $1.90 | 200k × $50 = $10.00 | $11.90 |
| Batch API, uncached (Fable 5 or 5.1) | 1M × $5 = $5.00 | 200k × $25 = $5.00 | $10.00 |
The honest read: output dominates Fable economics. In the uncached case, 200k output tokens cost the same as 1M input tokens. Caching attacks input, so it only halves a bill where input is large and reused; if your workload is output-heavy (long generations, verbose agents), the Batch API, which discounts output too, or downtiering to Opus 5.5 saves more. There is one more wrinkle that quietly inflates every Fable estimate, covered next.
Claude Opus 4.7 and later models, including Fable 5, 5.1 and 5.2, use a newer tokenizer that Anthropic says produces approximately 30% more tokens for the same text than older models. A naive per-token price comparison against an older model therefore understates real Fable spend. When you forecast, compare on the same text, not just the same token count, or you will under-budget.
Is Claude Fable 5 cheaper than Opus 4.8 or Opus 5.5? (side-by-side cost)
No. Fable 5.2 costs exactly 2× the current Opus 5.5 on base input, output, cache writes and batch rates, and the cache read is at parity: both charge $0.20. Fable 5.1 still costs 2.5× Opus 5.5, and Fable 5 costs exactly 2× the legacy Opus 4.8 on every row. Here is the comparison dollar for dollar.
| Operation (per MTok) | Fable 5.2 | Fable 5.1 | Opus 5.5 | Fable 5 | Opus 4.8 |
|---|---|---|---|---|---|
| Base input | $8.00 | $10.00 | $4.00 | $10.00 | $5.00 |
| Base output | $40.00 | $50.00 | $20.00 | $50.00 | $25.00 |
| Cache hit / read | $0.20 | $0.25 | $0.20 | $1.00 | $0.50 |
| Batch input | $4.00 | $5.00 | $2.00 | $5.00 | $2.50 |
| Batch output | $20.00 | $25.00 | $10.00 | $25.00 | $12.50 |
For a heavily cached agent loop, Fable 5.2’s premium over Opus 5.5 on input falls well below 2×, because cached reads cost the same, but output is still 2×, and output usually dominates. Anthropic’s own guidance is to start with Opus 5.5 for most workloads and use Fable 5.2 for demanding reasoning and long-horizon agentic work, or when your evals on Opus 5.5 at higher effort still fall short. The practical rule: if Opus 5.5 clears your quality bar on a task, it halves the per-token bill. Anthropic also claims Fable 5.2 at lower effort can match Opus 5.5 on some tasks for less cost per task; our Fable 5.2 guide covers those claims, which are worth testing on your own evals. Reserve Fable for work where the extra capability measurably changes the outcome, and route the rest to Opus 5.5 or a cheaper provider entirely; see our best inference APIs comparison for 2026 for non-Anthropic options.
How does Fable 5 draw down a subscription plan now?
As of November 12, 2026, Anthropic’s help center says Fable 5.2, Fable 5.1 and Fable 5 are available on all paid plans, but on different terms, and not on Free. On Max and Ultra, and on premium seats of Team and legacy seat-based Enterprise plans, Fable is a standard part of the plan: you can spend up to 50% of your weekly usage limit on Fable at no extra cost. On Pro and on standard Team and seat-based Enterprise seats, Fable isn’t included in plan limits and runs on pay-as-you-go usage credits, billed at standard API rates. On Enterprise, Fable 5.2 stays off until an admin enables it.
| Plan | Price | Fable 5.2, 5.1 and 5 access |
|---|---|---|
| Pro | $20/month, or $17/month billed annually | Usage credits at API rates ($8 / $40 on Fable 5.2); launch offer: Fable inside plan limits, up to 50% of weekly usage, through November 16, 2026, plus one Fable limit reset |
| Max | $100/month (5× Pro usage) | Included, up to 50% of the weekly limit; launch offer: 1 to 3 Fable limit resets, valid until March 8, 2027 |
| Ultra | $200/month (10× Pro weekly usage) | Included, up to 50% of the weekly limit; launch offer: 1 to 3 Fable limit resets, valid until March 8, 2027 |
| Team, standard seat | $20/seat annual, $25 monthly | Usage credits, if the organization enables them; same launch-week offer as Pro |
| Team, premium seat | $100/seat annual, $125 monthly | Included, up to 50% of the weekly limit; launch offer: 1 to 3 Fable limit resets |
| Usage-based Enterprise / API | Seat price plus usage at API rates | Standard API rates; on Enterprise, Fable 5.2 is off until an admin enables it |
The 50% is a share of your existing weekly limit, not extra allowance: other models draw from the same limit, and Anthropic says Fable models use it faster than other Claude models. Anthropic does not publish a burn multiplier. When you hit the Fable cap on Max or Ultra, you can keep using Fable with usage credits or switch to another model within your plan limits. Max 20x was renamed Ultra on October 30, 2026; our Claude Ultra plan explainer covers its weekly limit and the removal of the five-hour cap for verified accounts.
How we got here: this guide first reported that Fable 5 left Pro, Max and Team after a June 9–22 free window. Anthropic’s help center now dates the end of the original Fable 5 promotion, which allowed up to 50% of weekly limits on Fable 5, to July 19, 2026. Fable 5 then moved to usage credits on Pro and Team standard seats, with a one-time credit for eligible accounts; Anthropic says there is no equivalent credit for Fable 5.1 or 5.2.
Fable 5.2 launch offers: under Anthropic’s launch-offers article, Pro and Team standard seats get Fable inside plan limits, up to 50% of weekly usage, through November 16, 2026 at 11:59 pm PT, plus one Fable limit reset that expires at the same time. Max, Ultra, Team premium and premium seats on legacy seat-based Enterprise get one to three Fable limit resets by subscription length (one under six months, two for six to 12 months, three for over 12 months), valid until March 8, 2027. Billing must have started before November 9, 2026 at 10:00 am PT. Free, self-serve and usage-based Enterprise, and standard seats on legacy seat-based Enterprise are excluded, and HIPAA-ready Enterprise plans don’t get resets.
Fable on a plan works across Claude on the web, desktop and mobile, Claude Code, Cowork and other Claude surfaces. Claude Code needs version 2.1.321 or later for Fable 5.2 (2.1.255 or later for Fable 5.1, 2.1.170 or later for Fable 5), and its fable alias now resolves to Fable 5.2, except in Claude apps gateway sessions, where it still means Fable 5. Usage credits are separate from the monthly API credits that Max, Ultra and Team include, which can’t be spent in the Claude apps or Claude Code.
Is Fable 5 worth it on a plan versus the API?
For heavy Fable users, the plan now matters more than the rate card. Anthropic’s list prices: Pro is $20/month (or $17/month billed annually), Max is $100/month, Ultra is $200/month, Team is $20–25 per standard seat or $100–125 per premium seat, and Enterprise is $20 per seat billed annually plus usage at API rates. On Max or Ultra, Fable costs nothing extra until you reach half your weekly limit. On Pro, outside the launch week, every Fable message draws usage credits at $8 / $40 per million tokens on Fable 5.2 ($10 / $50 on 5.1 and 5).
The practical framing: if your Fable use is occasional and high-value, Pro plus credits is fine. If you reach for Fable every day, compare your monthly credit spend with the $80 step from Pro to Max, where Fable becomes part of the plan. If you run Fable heavily and predictably through code, model your monthly token volume at the API rate card above and compare it with Opus 5.5 at half the per-token price. For agent and Claude Code workloads specifically, where token volume balloons with tool loops and long context, the caching discipline from the rate-card section matters more than the plan choice; if the cost pushes you to weigh alternatives, see our 2026 comparison of AI coding assistants.
What are the API-side rate limits, separate from plan allowances?
API rate limits are a different ceiling from subscription plan allowances and are organized into usage tiers (Tier 1–4 plus Enterprise) that raise your requests-per-minute and tokens-per-minute as your spend history grows. This is the limit that bites production developers, not Claude.ai subscribers: it caps throughput, not monthly dollars. The structure mirrors what we documented for other inference providers in the NVIDIA NIM pricing and limits guide; rate limits, not just per-token price, often decide whether a workflow is actually shippable.
Because tier ceilings depend on your account history and can change, treat them as account-specific and confirm yours in the console rather than hard-coding a number. The engineering takeaway is the same as for any high-volume API: design for backoff, cap concurrency, and keep a cheaper fallback model wired in so a throughput cap doesn’t stall a run mid-workflow. For refusals specifically, Anthropic’s server-side fallback for Fable 5.2 can retry on Opus 4.8 or Opus 5; the fallback model can’t read Fable 5.2’s thinking blocks, so earlier reasoning is dropped.
FAQ
How much does Claude Fable 5 cost per message?
There is no fixed per-message price. Fable 5.2 bills per token at $8 per million input tokens and $40 per million output tokens; Fable 5.1 and Fable 5 bill $10 and $50. A short exchange of a few thousand tokens costs fractions of a cent; a long agent run with large context and verbose output can cost dollars. Use the worked example (1M input + 200k output = $16 uncached on Fable 5.2) to scale to your own message sizes.
Is Claude Fable 5 cheaper in batch mode?
Yes. The Batch API applies a flat 50% discount in both directions, taking Fable 5.2 to $4 per million input tokens and $20 per million output tokens, and Fable 5.1 and Fable 5 to $5 and $25. Batch is for asynchronous, non-real-time work like evals, bulk classification or offline summarization. Because it discounts output too, batch is often a bigger saving than prompt caching for output-heavy jobs.
Is Fable included in my Claude plan?
On Max and Ultra, and on premium Team seats, Fable 5.2, Fable 5.1 and Fable 5 are a standard part of the plan: you can use up to 50% of your weekly usage limit on them at no extra cost. On Pro and standard Team seats, Fable isn’t included in plan limits and runs on usage credits billed at standard API rates ($8/$40 per million tokens on Fable 5.2), except during the Fable 5.2 launch week through November 16, 2026. On Enterprise, an admin must enable Fable 5.2.
How does Fable 5 burn down my plan limit?
Anthropic says Fable models draw from your plan’s regular weekly limit and use it faster than other Claude models, but it does not publish a multiplier. On Max and Ultra you can spend up to 50% of the weekly limit on Fable. After that, you can keep using Fable with usage credits or switch to another model within your plan limits.
Is Fable 5 cheaper than Opus?
No. Fable 5.2 costs exactly 2x Opus 5.5 on input and output ($8 vs $4 input, $40 vs $20 output), and both charge $0.20 for cache reads. Fable 5.1 costs 2.5x Opus 5.5. The legacy pair keeps a flat 2x gap: Fable 5 at $10/$50 versus Opus 4.8 at $5/$25. If Opus 5.5 clears your quality bar, it halves the per-token bill.
Does prompt caching make Fable 5 meaningfully cheaper?
Yes, for reused context. Cache hits cost $0.20 per million tokens on Fable 5.2, 97.5% off the $8 base input, $0.25 on Fable 5.1 and $1 on Fable 5. The 5-minute cache write costs 1.25x base and pays off after one reuse; the 1-hour cache writes at 2x and pays off after two reuses. For long system prompts and agent loops that re-send the same context, caching is close to essential on an $8-input model.
Did Fable 5 come back to standard Claude plans?
Yes, on Max, Ultra and premium Team seats, where Fable 5.2, Fable 5.1 and Fable 5 are now a standard part of the plan at up to 50% of the weekly limit. Anthropic’s help center says the original Fable 5 promotion ended on July 19, 2026. Pro and standard Team seats still use usage credits for Fable.
Bibliography (19 sources)
Sources prioritise Anthropic primary documentation: the API pricing and model pages for rates, and the Claude help center and pricing page for plan rules. Vendor positioning claims are attributed to Anthropic. Earlier versions of this page relied on third-party developer guides for a reported plan burn multiplier; that figure was removed in the November 2026 update because Anthropic does not publish one. Links accessed November 12, 2026.
- Anthropic / Claude Platform Docs — Pricing (canonical rate card). platform.claude.com
- Anthropic — Introducing Claude Fable 5.2 and Claude Mythos 5.2 (announcement, November 9, 2026). anthropic.com
- Anthropic / Claude Platform Docs — What’s new in Claude Fable 5.2 (pricing, fast mode, breaking changes). platform.claude.com
- Anthropic / Claude Platform Docs — Claude Fable 5.1 overview. platform.claude.com
- Anthropic / Claude Platform Docs — What’s new in Claude Fable 5.1 (pricing, retention, breaking changes). platform.claude.com
- Anthropic — Claude Fable 5.1 and Claude Mythos 5.1 (announcement). anthropic.com
- Anthropic — Claude Fable 5 and Claude Mythos 5 (announcement). anthropic.com
- Anthropic — Pricing (subscription plans). claude.com/pricing
- Claude Help Center — Claude Fable models on your plan. support.claude.com
- Claude Help Center — Claude Fable 5.2 launch offers. support.claude.com
- Claude Help Center — Manage usage credits for paid Claude plans. support.claude.com
- Claude Help Center — What is the Ultra plan?. support.claude.com
- Anthropic / Claude Platform Docs — Prompt caching. platform.claude.com
- Anthropic / Claude Platform Docs — Batch processing. platform.claude.com
- Anthropic / Claude Platform Docs — Context windows. platform.claude.com
- Anthropic / Claude Platform Docs — Rate limits. platform.claude.com
- Claude Code Docs — Model configuration (fable alias, minimum version). code.claude.com
- Anthropic / Claude Platform Docs — Refusals and fallback. platform.claude.com
- TechCrunch — Anthropic’s Claude Fable 5 is a version of Mythos the public can access today (June 2026 launch coverage). techcrunch.com
