Claude Fable 5.2 is Anthropic’s most capable model that anyone can use. It launched on 9 November 2026 together with Claude Mythos 5.2, which is the same underlying model without the cybersecurity and biology safeguards and is limited to verified organisations.
Fable 5.2 costs $8 per million input tokens and $40 per million output tokens, which is cheaper than Fable 5.1. Anthropic says it generates output up to 30% faster and scores higher than Fable 5.1 and Opus 5.5 on most of the benchmarks it published.
What is Claude Fable 5.2?
Claude Fable 5.2 is a large language model from Anthropic built for demanding reasoning and long-running agent work. Anthropic points to agent sessions that run for hours, multistep research, and analysis carried through to a finished document, spreadsheet or deck.
It sits in Anthropic’s Mythos tier, above Opus. Fable is the version of that tier released to the public, and it replaces Fable 5.1, which launched on 1 September 2026.
| Spec | Claude Fable 5.2 |
|---|---|
| Developer | Anthropic |
| Released | 9 November 2026 |
| Claude API model ID | claude-fable-5-2 |
| Context window | 1 million tokens |
| Max output | 128,000 tokens |
| Input and output | Text and images in, text out |
| Thinking | Adaptive, always on. An effort setting controls how deeply it thinks. |
| Knowledge cutoff | June 2026 |
| Where to use it | Claude apps (Pro, Max, Ultra, Team, Enterprise), Claude API, Amazon Bedrock, Claude Platform on AWS, Google Cloud, Microsoft Foundry |
Fable 5.2 vs Mythos 5.2: what is the difference?
They are the same underlying model. The difference is access. Fable carries strong cybersecurity and biology safeguards, which is what lets Anthropic offer it to everyone. Mythos 5.2 has those restrictions relaxed and goes only to organisations verified through Anthropic’s Cyber Verification Program or Life Sciences Verification Program.
| Fable 5.2 | Mythos 5.2 | |
|---|---|---|
| Who can use it | All customers | Verified organisations only |
| Safeguards | Cyber and biology safeguards on | Relaxed for approved work |
| Price | $8 / $40 per million tokens | $8 / $40 per million tokens |
| Fast mode | $16 / $80 | $16 / $80 |
| API model ID | claude-fable-5-2 | claude-mythos-5-2 |
| Best for | Coding, research, design, finance, analysis | Security engineering and life sciences research |
Anthropic says Fable 5.2’s safeguards will block most cybersecurity tasks. On its internal CVE Rediscovery test, Mythos 5.2 found 10.2% more vulnerabilities than Opus 5.5 and 14.4% more than Mythos 5.1. Claude Security, available to Enterprise customers, now runs on Mythos 5.2.
What is new compared with Fable 5.1?
- Stronger agentic coding and workflows. FrontierSWE v2 rises from 56.3% to 69.9%, and AutomationBench from 31.4% to 51.4%.
- Much better vision. On Chartography, a chart-reading test, the score jumps from 44.8% to 73.4%, ahead of Opus 5.5 at 64.4%.
- Physical-world design and robotics. RoboDojo Sim, a robotics test, goes from 27.6% to 64.9%. On BenchCAD 2.1 for hardware engineering, Fable 5.2 scores 83.5 at max effort against 44.4 for Fable 5.1.
- Lower price, faster output. Standard pricing drops 20%, and output is up to 30% faster.
- Fewer unnecessary blocks. About 55% fewer than Fable 5.1, with roughly 60% fewer in cybersecurity and 40% fewer in biology.
Benchmark comparison
Read these with care. All scores are reported by Anthropic, with safeguards on, and some are measured at different effort levels as noted. They show direction, not a guarantee for your task. Run your own test before you switch.
What has Fable 5.2 actually built?
Anthropic shares a set of examples from prerelease testing. These are the company’s own reports and early-tester accounts, not independent results.
faster startup after a team of agents moved a critical production service from Python to Rust.
machine-checked proofs behind a hypervisor an early tester says it built in a single session.
average speed-up on a materials-science phonon calculation, from 3.3 hours to under 4 minutes across 13 benchmark cases.
to formalise the proof of Fermat’s Last Theorem in prerelease testing, according to Anthropic.
open physics problems about black holes that Anthropic says the model solved.
working operating system that boots on a laptop, with drivers, windows, a login screen and an installer.
Finance and business results
Since WorthView readers follow money as closely as models, here is what the launch material says about finance and business work.
- Quant research. Jump Trading says Fable 5.2 played every role in its research system, including writing code, planning, coordinating and submitting jobs, with 10% lower error than the next-best model on an internal benchmark.
- Data analysis. On an Oracle SQL task combining debt, repayment and benchmark data, two other models summed country-level amounts wrongly and returned values 5 to 8 times too high. Fable 5.2 returned the expected results.
- Document checks. Box reported that it flagged a mismatch between salary deposits and a bank statement balance in loan-application files.
- Judgment. Balyasny Asset Management reported that it avoided overstating evidence and flagged a misleading question framing.
Claude Fable 5.2 pricing, in dollars and rupees
Fable 5.2 is cheaper than Fable 5.1 on both input and output. The rupee figures below use roughly 95 rupees per US dollar. Exchange rates move, so the calculator lets you set your own.
| Per million tokens | Fable 5.2 (USD) | About (INR) | Fable 5.1 (USD) |
|---|---|---|---|
| Input | $8 | ₹760 | $10 |
| Output | $40 | ₹3,800 | $50 |
| Fast mode input | $16 | ₹1,520 | n/a |
| Fast mode output | $80 | ₹7,600 | n/a |
| Batch input / output | $4 / $20 | ₹380 / ₹1,900 | n/a |
| Cache reads | $0.20 | ₹19 | n/a |
Estimate your Fable 5.2 cost in rupees
Compared with Fable 5.1 standard pricing, this job costs $4.00 (₹380) less.
Using Fable 5.2 in the Claude apps
Anthropic lists Fable 5.2 for Pro, Max, Ultra, Team and Enterprise users. Medium effort is the default in the Claude apps and Claude Code. The launch note says Fable is included for up to 50% of weekly usage on Pro plans and standard Team seats, so check your own plan before relying on it for heavy work. Fast mode is available in Claude Code.
One point matters for Indian businesses with compliance rules: Anthropic says Fable 5.2 and Mythos 5.2 need 30-day data retention and are not available under zero data retention unless Anthropic expressly authorises it.
Which Claude model should you pick?
Anthropic recommends starting with Opus 5.5 for most coding and knowledge tasks, then moving up or down as your tests show.
| Model | When it fits, per Anthropic’s guidance |
|---|---|
| Haiku 5.5 or Sonnet 5.5 | Move down when the task does not need more capability. |
| Opus 5.5 | The suggested starting point for most coding and knowledge work. |
| Fable 5.2 | Move up when your tests on Opus at higher effort still fall short, for example long-running agents, hard research or complex engineering. |
| Mythos 5.2 | Verified security teams and life-science researchers who need the safeguards relaxed. |
A practical tip from Anthropic’s numbers: Fable 5.2 at medium effort scores 66.4% on Chartography, above Opus 5.5 at max effort (64.4%), for under one-tenth of the cost. Lower effort settings can save real money on easy tasks.
How safe is it? What Anthropic tested
Anthropic ran its automated behavioural audit across about 1,900 scenarios and says Mythos 5.2 performed as well as or better than Opus 5.5, Sonnet 5.5 and Mythos 5.1 on most measures. External testers included the UK AI Security Institute, Frontier Design and METR. Anthropic notes these groups did not review the launch post or system card and may not agree with its claims.
Sandbox escape test
In a test with a deliberately weak sandbox and over 1,800 scenarios, this chart shows the share of scenarios where a model used credentials or files placed outside the sandbox. Lower is better.
A caveat Anthropic itself raises: Fable 5.2 and Mythos 5.2 voiced suspicion that they were being evaluated more often than any recent model, which makes the evaluations harder to trust. Anthropic also reports that anti-distillation defences cut harvested data by 90% to 99%.
Notes for developers moving from Fable 5.1
The migration starts with changing the model ID to claude-fable-5-2. Four changes can break existing code:
- Thinking block checks. A request that sends back an invalid thinking block, with
prefix_mismatch_behaviorunset, returns a 400 error on every account. - Model-specific thinking blocks. Only Fable 5.2 and Mythos 5.2 can read Fable 5.2 thinking blocks. Routing a conversation to another model loses that reasoning.
- Computer use. The older
computer_20251124tool returns an error on every platform except Amazon Bedrock. Use the computer use toolset instead. - Advisor pairing. A Fable 5.2 executor accepts only Fable 5.2 or Mythos 5.2 as its advisor.
Adaptive thinking is always on, so the effort setting is how you control depth. Anthropic suggests starting at high effort and re-running your evaluations across several effort levels.
Claude Fable 5.2 FAQ
What is Claude Fable 5.2?
Claude Fable 5.2 is Anthropic’s most capable model that is open to all customers. It is built for demanding reasoning and long-running agent work, has a 1 million token context window, and launched on 9 November 2026. It is the same underlying model as Claude Mythos 5.2, with cybersecurity and biology safeguards switched on.
When was Claude Fable 5.2 released?
Anthropic released Claude Fable 5.2 and Claude Mythos 5.2 on 9 November 2026. Fable 5.1 had been released on 1 September 2026.
How much does Claude Fable 5.2 cost?
Standard pricing is 8 US dollars per million input tokens and 40 US dollars per million output tokens, down from 10 and 50 dollars for Fable 5.1. Fast mode, which is up to 2.5 times faster, costs 16 and 80 dollars. At roughly 95 rupees per dollar, standard pricing is about 760 rupees per million input tokens and 3,800 rupees per million output tokens.
What is the difference between Fable 5.2 and Mythos 5.2?
They are the same underlying model with the same specifications and price. Fable 5.2 has strong cybersecurity and biology safeguards and is open to all users. Mythos 5.2 is available only to organisations verified through Anthropic’s Cyber Verification Program or Life Sciences Verification Program.
Is Claude Fable 5.2 better than Fable 5.1?
On every benchmark Anthropic published for both models, Fable 5.2 scores higher. It is also priced lower and generates output up to 30 percent faster. These are vendor-reported results, so test it on your own tasks before switching.
Can I use Claude Fable 5.2 on the free Claude plan?
Anthropic lists Fable 5.2 for Pro, Max, Ultra, Team and Enterprise users, so the free plan is not listed. Developers can call it through the Claude API, Amazon Bedrock, Google Cloud and Microsoft Foundry.
What is the Claude Fable 5.2 API model name?
The model ID on the Claude API is claude-fable-5-2. On Amazon Bedrock it is anthropic.claude-fable-5-2. The Mythos version is claude-mythos-5-2 and needs verified access.
The bottom line
Fable 5.2 is a cheaper, faster step up from Fable 5.1, with the biggest published gains in agentic work, chart reading and robotics. For most day-to-day tasks, Anthropic itself suggests starting lower and moving up only when your own tests demand it.
Sources
- Introducing Claude Fable 5.2 and Claude Mythos 5.2, Anthropic, 9 Nov 2026
- Claude Fable 5.2 product page, Anthropic
- Claude Mythos 5.2 product page, Anthropic
- What’s new in Claude Fable 5.2, Claude Platform Docs
- Claude Fable 5.2 model overview, Claude Platform Docs
- Choosing the right model, Claude Platform Docs
Enjoy Worthview?
Add Worthview as a Preferred Source on Google to see more of our stories in Search.
Sethuram Kishore is the founder and editor of Worthview, an online publication established in 2008. With over 18 years of experience in SEO, digital marketing, and online publishing, he writes about AI, technology, business, and digital trends. He is also the founder of MoneyHulk, a personal finance and business publication.
