Claude Fable 5.2 and Mythos 5.2: pricing, access and launch offers

Anthropic launched Claude Fable 5.2 and Claude Mythos 5.2 on November 9, 2026, lowering standard API input and output list prices from $10 and $50 to $8 and $40 per million tokens. Anthropic says they share the same underlying model, and that Mythos access remains limited to organizations verified through programs such as the Cyber Verification Program. Fable 5.2 is listed on the Claude API and on partner platforms including Amazon Web Services, Google Cloud and Microsoft Foundry. See the announcement and Toolvel’s pricing changes log.

API price comparison

All figures below are USD per million tokens before tax. Paired figures mean input/output. On some partner platforms, including Google-managed Google Cloud, the cloud provider sets the prices, so check that platform’s list rather than assuming first-party rates.

Price category	Fable 5.2	Mythos 5.2	Fable 5.1	Mythos 5.1
Standard input	$8	$8	$10	$10
Standard output	$40	$40	$50	$50
Cache read	$0.20	$0.20	$0.25	$0.25
5-minute cache write	$10	$10	$12.50	$12.50
1-hour cache write	$16	$16	$20	$20
Batch input/output	$4/$20	$4/$20	$5/$25	$5/$25
Fast mode input/output	$16/$80	$16/$80	Not listed on the pages Toolvel checked	Not listed on the pages Toolvel checked

Toolvel calculation: standard input and output list prices are 20% below the 5.1 rates, because (10 − 8) / 10 = 0.20 and (50 − 40) / 50 = 0.20. Anthropic published the dollar amounts, not this percentage.

API access and billing conditions

Use claude-fable-5-2, or claude-mythos-5-2 with verification. Both have a 1M-token context window (the default and the maximum), 128k maximum output tokens and always-on adaptive thinking. Official pricing says Claude 4.6 and later models, except Claude Haiku 5.5, include the full 1M context window at standard per-token rates. See the model overview and API pricing.

Fast mode costs $16/$80. Anthropic says it delivers up to 2.5 times higher output tokens per second. On the Claude API, set speed: "fast" on the request, with no beta header. Official availability lists also include Claude Code, Claude Models on Google Cloud (Anthropic managed and Google managed), and Microsoft Foundry for deployments hosted on Anthropic, where the Foundry model ID is claude-fable-5-2-fast. Mythos fast mode is limited to Verification Program organizations. Fast mode is not available with the Batch API. Prompt-caching and data-residency multipliers still apply on top of fast mode prices. US-only inference adds a 1.1x multiplier across token categories, including input, output, cache writes and cache reads. Sources: the release notes, What’s new and pricing.

Both models require 30-day data retention for safety monitoring, like Fable 5.1 and Mythos 5.1, and are not available under zero data retention unless Anthropic expressly authorizes it. Anthropic’s announcement describes Enterprise Frontier Safeguards as a private beta and says eligible customers can continue using Fable 5.2 and Mythos 5.2 with zero data retention before implementing it. The Fable product page says API customers must configure the Fallback API for safeguard routing. The permitted fallback targets listed for Fable 5.2 are Claude Opus 4.8 and Claude Opus 5.

API migration changes

The What’s new in Claude Fable 5.2 guide names four changes from Fable 5.1 that can break existing code:

• Prefix check: Fable 5.2 and Mythos 5.2 enforce the prefix check for every account, whatever its age. If your code edits the system prompt, tools or an earlier message and then sends back a 5.2 thinking block, the request returns a 400 error. The documented error message is “The block is bound to a different conversation.” To drop the affected blocks and continue, the guide says to send the thinking-binding-controls-2026-08-01 beta header and set thinking.block_binding.prefix_mismatch_behavior to "drop_block". Nothing changes if Claude Code, claude.ai, Claude Managed Agents or the Claude Agent SDK builds your requests.
• Model switching: on the Claude API and Google-managed Google Cloud, moving onto Fable 5.2 or Mythos 5.2 keeps reasoning from other models, including Claude Opus 5.5, Claude Sonnet 5.5 and Claude Haiku 5.5. Switching away drops the 5.2 thinking blocks, and the request still succeeds. Thinking blocks that these models produce work only in the account that produced them, or in a linked account.
• Computer use: use the computer_toolset_20260801 toolset. A request that declares the earlier computer_20251124 tool returns a 400 error on every platform except Amazon Bedrock, where the earlier tool still works.
• Advisor tool (beta): a Fable 5.2 or Mythos 5.2 executor accepts only Claude Fable 5.2 or Claude Mythos 5.2 as its advisor. Any other advisor returns a 400 error. Every executor model accepts Fable 5.2 as its advisor.
• Request settings: omit the thinking field or send {"type": "adaptive"}. thinking type enabled with budget_tokens, thinking type disabled, and tool_choice types any or tool return 400 errors. The API default effort is high. Official docs also say to omit temperature, top_p and top_k. In an organization with customer-managed encryption keys, structured outputs are not available for these models.

Claude Code and paid-plan access

Claude’s pricing page lists no Fable access on the Free plan. After the launch offers end, Pro, standard Team and standard seat-based Enterprise use usage credits for Fable models. Max, Ultra, premium Team and premium seat-based Enterprise include Fable for up to 50% of weekly limits. That 50% is part of the weekly limit, not extra. Claude Code is included in all paid plans and shares the same usage limits. Anthropic’s plan-access guide also says Fable models use those limits faster than other Claude models. Toolvel’s Claude Code pricing hub covers the subscription and billing distinction.

Fable 5.2 requires Claude Code version 2.1.321 or later, according to both the Help Center and the Claude Code model configuration page. Select it with /model fable, or launch with claude --model fable. Unless you set ANTHROPIC_DEFAULT_FABLE_MODEL, the fable alias resolves to Fable 5.2 except in Claude apps gateway sessions, where it resolves to Fable 5; use /model claude-fable-5-2 there. No Fable model is the account-type default on any plan or provider. Claude Code’s default effort for Fable 5.2 is medium, which differs from the Claude API default of high.

A saved claude-fable-5-1 user setting can move to the fable alias on Pro, Max and Ultra the first time you run Claude Code 2.1.321 or later; on Team and Enterprise, the move waits for organization settings and may come later or not at all. On Enterprise plans, Fable 5.2 stays off until an admin enables it. Seven days after launch, it is added to every role that has any Fable model enabled, and any default set to a Fable model moves to Fable 5.2, unless an admin opts out first. See the plan-access guide.

Launch offers and reset deadlines

The launch-offer terms start November 9, 2026 and cover Fable 5.2, Fable 5.1 and Fable 5:

• Pro plans and standard Team seats get all Fable models included for up to 50% of weekly limits through November 16, 2026 at 11:59 pm PT, plus one Fable limit reset that expires at the same time. After you use that 50%, or once the week ends, Fable usage is billed to usage credits if you have them turned on. Unused allowance does not carry over.
• Max plans, Ultra plans, premium Team seats and premium seats on legacy seat-based Enterprise plans receive one Fable limit reset if the paid subscription is under six months old, two if it is six to 12 months, or three if it is more than 12 months. These resets expire March 8, 2027 at 11:59 pm PT.

Resets require paid billing in place before November 9, 2026 at 10:00 am PT. For Team and Enterprise seats, the organization’s billing must also have been in place by then, and seat holders must have joined the organization before then. If a subscription started before the cutoff but is still in a free trial, the reset arrives once it converts to paid; self-serve Team trials count as paid. The offers are not available on the Free plan, standard seats on legacy seat-based Enterprise plans, self-serve Enterprise plans or usage-based Enterprise plans. Fable limit resets are also not available for HIPAA-ready Enterprise plans.

A reset restores weekly usage spent on Fable models only; usage spent on other models stays where it is. If the account has a five-hour limit, the reset refills that whole window, including usage on other models. Verified Ultra accounts have no five-hour limit, so a reset only affects the weekly limit. A reset does not raise your limits, cannot be undone, and is no longer available if you cancel or downgrade before using it. To redeem one, go to Settings > Usage and, under Limit resets, select “Reset for free.”

What official pages do not state yet

The official pages Toolvel read do not state absolute weekly or five-hour Fable usage amounts in tokens or messages. Their fast-mode availability lists do not name Amazon Bedrock; that is not the same as a confirmation that Bedrock has no fast mode. They give no clock time for the November 9 launch itself, beyond the separate 10:00 am PT billing cutoff for launch-offer resets, and they do not name a calendar date for the Enterprise “seven days after launch” admin rule. They also do not say whether unverified accounts see Mythos 5.2 in the standard Claude Code model picker. See Claude pricing, the announcement and the plan-access guide.

Official sources

• Introducing Claude Fable 5.2 and Claude Mythos 5.2
• Claude Platform release notes
• What’s new in Claude Fable 5.2
• Claude Fable 5.2 model overview
• Claude API pricing
• Claude Fable product page
• Claude plans and pricing
• Claude Fable 5.2 launch offers
• Claude Fable models on your plan
• Claude Code model configuration

Source check: November 11, 2026 (PT). Toolvel reviewed official documents only and did not run usage tests. Prices and plan terms can change. Read how Toolvel checks pricing.