Claude Fable 5.2 vs Fable 5.1: What Changed and Should You Switch?

Fable 5.2 is 20% cheaper on every API rate

Anthropic released Fable 5.2 on November 9, 2026, and cut the price at the same time. Read each row as the same charge on both models; every standard rate falls by a fifth.

As an illustration, a Claude Code turn that sends 40,000 input tokens and returns 2,000 output tokens without cache reads costs $0.32 + $0.08 = $0.40 on Fable 5.2 and $0.40 + $0.10 = $0.50 on Fable 5.1. The turn size is ToolColumn's example, not an Anthropic figure. Per finished task the saving can be larger: Anthropic says Fable 5.2 at medium effort, the default in Claude apps and Claude Code, completes hard tasks with fewer tokens than Fable 5.1.

Fast mode is new for Fable. It speeds up output by up to 2.5 times at double the standard rate, in Claude Code, on Anthropic's API, on Google Cloud and on Microsoft Foundry.

Anthropic's figures show the biggest gains on coding and long tasks

The table takes the rows of Anthropic's launch table that report both models, with safeguards on and each model at its highest effort unless noted. These are Anthropic's own results, not independent tests.

Robot control, chart reading, long knowledge work and agentic coding move most, by 14 to 37 points; Humanity's Last Exam moves about nine and computer use about six. Fable 5.1 had trailed Opus 5.5 at that model's launch, and Fable 5.2 now scores above Opus 5.5 on every one of these rows, which Claude Opus 5.5 vs Fable 5.2 vs Sonnet 5.5 compares.

Plan access did not change, but Claude Code needs an update

Fable 5.2, 5.1 and 5 follow one set of rules on Claude plans. On Max, Ultra and Team Premium seats they share up to 50% of the weekly limit and use it faster than other models. On Pro and Team Standard seats they run on usage credits at API rates. Free has none of them. For the launch week, Pro and Team Standard seats get Fable inside the plan up to 50% of weekly limits, from November 9 through November 16, 2026 at 11:59 pm PT, plus one Fable limit reset; Max, Ultra and Premium seats get one to three resets, depending on how long the account has paid, which expire on March 8, 2027.

In Claude Code, Fable 5.2 requires version 2.1.321 or later. The fable alias now resolves to Fable 5.2, except in Claude apps gateway sessions, where it still means Fable 5. To keep Fable 5.1, select it by its model ID, claude-fable-5-1. On Enterprise, Fable 5.2 stays off until an admin enables it; seven days after launch Anthropic adds it to every role that already has a Fable model and moves Fable defaults to 5.2, unless an admin opts out first.

API code written for Fable 5.1 can fail on Fable 5.2

Changing the model ID to claude-fable-5-2 is not the whole migration. Anthropic lists four changes that can break existing code:

• Thinking blocks are checked on every account. Sending back a thinking block that does not belong to the conversation returns a 400 error unless prefix_mismatch_behavior is set. On Fable 5.1 the check applied only to accounts created on or after August 31, 2026.
• Thinking blocks stay with Fable 5.2. Only Fable 5.2 and Mythos 5.2 read them; a turn that switches to another model, Fable 5.1 included, drops them.
• The old computer-use tool is refused. computer_20251124 returns a 400 error everywhere except Amazon Bedrock; use computer_toolset_20260801.
• The advisor must match. A Fable 5.2 or Mythos 5.2 executor accepts only Fable 5.2 or Mythos 5.2 as its advisor.

Both models keep a 1-million-token context window and 128,000 tokens of output, and both require 30-day data retention, with zero data retention only where Anthropic expressly authorizes it. Fable 5.1 stays available as a legacy model, with retirement not sooner than September 1, 2027, so an integration can stay on it while the four changes are tested.

Mythos 5.2 replaces Mythos 5.1 for verified organizations

Mythos 5.2 is the same model as Fable 5.2 without some of its cybersecurity and biology safeguards, at the same $8/$40. Access still goes through Anthropic's Cyber Verification Program or Life Sciences Verification Program, and its fast mode is limited to members of those programs. On Anthropic's CVE Rediscovery test, Mythos 5.2 found 14.4% more vulnerabilities than Mythos 5.1. Claude Enterprise customers outside the programs reach it only inside Claude Security, Anthropic's codebase scanner.