Claude Fable 5.2 Is Here: What's New, What It Costs and Who Gets Mythos
Anthropic has launched Claude Fable 5.2 and Mythos 5.2. Here is what changed, the new API price, who can get Mythos, and what it means for users in Nigeria.

Anthropic released Claude Fable 5.2 and Claude Mythos 5.2 on 9 November 2026. Claude Fable 5.2 is the public version, available to everyone on Anthropic's platforms with strong safeguards that block high-risk cybersecurity and biology requests. Mythos 5.2 is the same underlying model without those blocks, and it is only for vetted organisations. Fable 5.2 is also cheaper than the version it replaces: $8 per million input tokens and $40 per million output tokens, down from $10 and $50.
This article covers what changed in Claude Fable 5.2, what Anthropic says it can do, what it costs (with naira estimates), who can actually get Mythos 5.2, and what it means for students, developers and small businesses in Nigeria. Where a claim comes only from Anthropic, we say so. Where it comes from news reports, we name them.
What Happened
Anthropic published its announcement, titled "Claude Fable 5.2 and Claude Mythos 5.2", on Monday 9 November 2026. Some news digests dated the launch 10 November, but Anthropic's own page carries the 9 November date. It describes Fable and Mythos as "models that are designed to tackle the most daunting technical challenges."
The short version:
- Claude Fable 5.2 is available "today on all platforms, including Amazon Web Services, Google Cloud, and Microsoft Azure", according to Anthropic. Developers call it with the model name
claude-fable-5-2. - Claude Mythos 5.2 is the same model with the cyber and biology capabilities left open. It is limited to organisations that pass Anthropic's verification checks.
- Price: $8 per million input tokens and $40 per million output tokens for standard use. That is 20% cheaper than Fable 5.1, which launched in September at $10 and $50, as the tech news site Implicator noted.
- Speed: Anthropic says outputs are "up to 30% faster" than Fable 5.1 and Mythos 5.1. There is also a separate fast mode that costs double.
If you have followed this model family, you will know it has had a bumpy year. Fable 5 launched in June as the first public Mythos-class model, and was briefly pulled worldwide after a US government directive before coming back in July. We covered that in what the Fable 5 launch meant and why Fable 5 came back. Version 5.2 is the second upgrade since then, and the first to bring a price cut.
Why It Matters
Three things make this release worth paying attention to, even if you never touch the API.
First, the price drop. Frontier models usually launch at a premium and stay there. A 20% cut on Anthropic's most capable public model lowers the cost of the heaviest coding and research work, and that tends to flow through to the apps and tools built on top of it.
Second, the safeguards are getting more precise. Anthropic says Fable 5.2 blocks about 55% fewer requests by mistake than Fable 5.1, including roughly 60% fewer false blocks in cybersecurity and 40% fewer in biology. For ordinary users, that should mean fewer frustrating refusals on harmless questions about, say, how a virus spreads or how to secure a home router.
Third, it shows where the industry is heading: long-running AI agents that work for hours or days. Anthropic's headline example is an engineer who asked Fable 5.2 to build a prototype operating system from an empty folder. According to Anthropic, "the model coordinated over 1,000 subagents, working around the clock for eight days." The company says the result boots on a laptop. That is Anthropic's own account and has not been independently reproduced, but it shows what the company is aiming at.
If you are trying to understand how close these systems are to "general" intelligence, our same-week explainer What Is AGI? is a calm place to start.
The Details
What's new in Claude Fable 5.2
Anthropic's announcement lists a long set of improvements over Fable 5.1. The main ones, all as reported by Anthropic:
- Coding: On FrontierCode 2.0, an agentic coding test, Fable 5.2 scored 49.1% at its highest effort setting, against 32.4% for Fable 5.1 and 45.5% for Claude Opus 5.5. On FrontierSWE v2 it scored 69.9%, up from 56.3%.
- Computer use: On OSWorld 2.1, which tests an AI's ability to operate a computer, it scored 89.3%, up from 83.5%.
- Research and reasoning: On Humanity's Last Exam with tools, it scored 74.2%, up from 65.6%.
- Efficiency: At medium effort, which is the default in the Claude apps and Claude Code, Anthropic says it completes tasks with fewer tokens than Fable 5.1.
- Honesty: Anthropic quotes Adhyyan Sekhsaria of Cognition saying: "Claude Fable 5.2 says it doesn't know things instead of hallucinating."
Anthropic also lists science examples. It says a team of about 1,000 Fable 5.2 agents formalised Fermat's Last Theorem in 30 hours, and that the model helped design three working electronic devices with a hardware partner. It also notes that rival models scored higher on some tests, including Google's Gemini 4 Carbon on an automation benchmark. These are company-run results, so treat them as claims rather than settled facts.
One point people miss: Anthropic does not say Fable 5.2 is the right model for everything. It recommends starting most coding and knowledge work on Claude Opus 5.5, moving up to Fable or Mythos for the largest tasks, and down to Sonnet 5.5 or Haiku 5.5 for simpler ones. Fable 5.2 can also act as an "Advisor" or "Orchestrator" that hands work to smaller, cheaper models.
What Claude Fable 5.2 costs
Here are Anthropic's published API prices, with naira estimates at roughly ₦1,400 to the dollar. Exchange rates move, so treat the naira column as a guide only.
| Option | Input (per million tokens) | Output (per million tokens) |
|---|---|---|
| Fable 5.2 standard | $8 (about ₦11,200) | $40 (about ₦56,000) |
| Fable 5.2 fast mode | $16 (about ₦22,400) | $80 (about ₦112,000) |
| Fable 5.1 (previous) | $10 (about ₦14,000) | $50 (about ₦70,000) |
Fast mode generates tokens up to 2.5 times faster, according to Anthropic. It is available in Claude Code, on the Claude Platform, through Google Cloud and through Microsoft Foundry.
For people on Claude subscriptions rather than the API, Anthropic says plans and seats that include Fable get "one to three Fable resets based on how long they've subscribed", and that "Fable is included for up to 50% of weekly usage limits for Pro plans and standard seats on the Team plan." Anthropic's page does not spell out access for every plan tier, so check the model menu in your own Claude account.
Who gets Mythos 5.2
This is the part that confuses most people. Fable 5.2 and Mythos 5.2 are the same model. The difference is the safety layer.
- Fable 5.2 runs with classifiers that block high-risk cyber activity, such as exploit generation and penetration testing, and most research-level biology tasks. When a request is blocked, Anthropic says, "users will instead receive a response from Opus 4.8 for cybersecurity and Opus 5 for biology." In other words, you still get an answer, just from an older, less capable model.
- Mythos 5.2 has those capabilities open. Cybersecurity professionals can apply through Anthropic's Cyber Verification Program, and life sciences researchers through its Life Sciences Verification Program. Claude Enterprise customers can also use Claude Security, which runs on the full model family including Mythos 5.2.
So, for almost everyone reading this, the answer to "can I get Mythos?" is no, unless you work for a verified security or life sciences organisation. That is by design.
Anthropic says Mythos 5.2 found 10.2% more known vulnerabilities than Opus 5.5 in an October 2026 test across 1,829 published security flaws. That kind of capability is exactly why it is restricted.
Safety testing
Anthropic says it rates the risk of unsafe behaviour in high-stakes situations as low, and that outside testers included the UK AI Security Institute and METR. It also reports one notable improvement: in a sandbox-escape test, Mythos 5.2 used credentials or files outside its sandbox in 3.1% of cases, compared with 21% for Mythos 5.1. Anthropic adds a frank caveat: both models sometimes suspected they were being tested, which it says makes evaluation harder.
What It Means for You
If you are a developer
The price cut is the practical headline. If you already use Fable 5.1 through the API, switching the model name to claude-fable-5-2 should cut your bill by about a fifth for the same token volume, and Anthropic says the model uses fewer tokens at medium effort too.
Here is an illustrative example. Chidi, an invented developer in Lagos, runs a contract-review tool for small law firms. His tool sends about 200,000 input tokens and gets back about 50,000 output tokens a day.
- On Fable 5.1: $2.00 input plus $2.50 output, so $4.50 a day, or about $135 for 30 days (about ₦189,000 at roughly ₦1,400 to the dollar).
- On Fable 5.2: $1.60 input plus $2.00 output, so $3.60 a day, or about $108 for 30 days (about ₦151,200).
That is a saving of about ₦37,800 a month before any token savings from the model being more efficient. Still, test before you switch everything. Anthropic recommends Opus 5.5 as the starting point for most work and Sonnet 5.5 or Haiku 5.5 for simpler jobs, and it says Fable 5.2 at high effort roughly matched Opus 5.5 on one coding test at about half the cost per task. The right choice depends on your workload, so compare real bills. Our guide to choosing an AI coding agent walks through that decision.
If you use the Claude app
If your plan includes Fable, you should see Fable 5.2 in the model picker. Remember the 50% cap: on Pro and standard Team seats, Fable can use up to half of your weekly limit, so save it for big jobs such as long reports, complex spreadsheets or debugging a stubborn codebase. For everyday writing and study help, a smaller model is usually enough and keeps your allowance for when you need it. If you are weighing Claude against ChatGPT or Gemini, our GPT-5 vs Claude vs Gemini decision guide compares them by task.
If you are a student or job seeker
Fable 5.2 is not a must-have for coursework or CV writing. Its strengths are long, technical, multi-step tasks. The more useful takeaway is the reduction in false refusals: if you study medicine, microbiology or computer security, you should hit fewer blocks on normal academic questions.
If you are in Nigeria
Nigeria is on Anthropic's list of supported countries, so you can sign up for Claude and the Claude API from here. The practical hurdles are payment and cost, not access.
- Payment: Claude subscriptions and API credits are billed in US dollars. Many Nigerian naira cards have low or changing limits on international spending, so payments can fail. Check your bank's current foreign-spend limit before you subscribe, or use a dollar card from a provider you trust.
- Cost control: API spending can climb quickly with a model this powerful. Set a monthly spending limit in your developer console and start with small tests.
- Cloud options: Because Fable 5.2 is also available through AWS, Google Cloud and Microsoft Azure, a Nigerian company that already pays one of those providers may find it easier to use Claude through its existing account.
Also This Week at Anthropic
The model launch was not the only news. On 10 November, Anthropic announced Claude Sheets and Claude Whiteboards in beta for all paid plans. Sheets builds and edits shared spreadsheets with formulas and drop-downs, and exports to .xlsx or CSV. Whiteboards turns rough plans into shared boards with timelines, owners and dates. On 12 November, it introduced Claude Meetings in beta, starting with Team and Enterprise plans. Claude joins Google Meet calls as a visible participant, answers questions and writes notes and action items. Anthropic says Google Meet is the only supported platform for now.
Anthropic also launched the roadshow for its stock market listing on 10 November, setting a price range of $70 to $78 per share, as RTTNews and other outlets reported. The shares are set to list on the Nasdaq under the ticker ANT. We covered the filing in Anthropic's S-1 and Nasdaq listing explained.
What to Watch Next
- Independent tests. Most performance figures so far come from Anthropic and its launch partners. Watch for independent benchmark results over the coming weeks.
- Rival pricing. Google released Gemini 4 Hydrogen in the same week at a much lower price for everyday agent work. Read our report on Gemini 4 Hydrogen to see how the two compare on cost.
- Usage limits. The Fable resets and the 50% cap are launch-period terms. Watch for changes once demand settles.
- Mythos access. Anthropic has expanded verification programmes for security and life sciences. Whether it opens similar routes for other fields is an open question.
- Enterprise Frontier Safeguards. Anthropic says a version of its safety monitoring that runs on customers' own infrastructure is in private beta. Wider availability would matter for banks and other regulated businesses.
Final Thoughts
Claude Fable 5.2 is an incremental but meaningful upgrade: cheaper, faster and, by Anthropic's account, better at long technical work, with fewer mistaken refusals. Mythos 5.2 remains a restricted tool for vetted security and science teams. For most readers in Nigeria, the sensible move is to try Fable 5.2 on one real, heavy task, compare it with a cheaper model, and only then decide whether it earns a place in your routine.
Frequently Asked Questions
What is Claude Fable 5.2?
Claude Fable 5.2 is Anthropic's most capable publicly available AI model, released on 9 November 2026. It is built for large coding, research and agent tasks, and runs with safeguards that block high-risk cybersecurity and biology requests.
How much does Claude Fable 5.2 cost?
Through the API, Anthropic charges $8 per million input tokens and $40 per million output tokens, about ₦11,200 and ₦56,000 at roughly ₦1,400 to the dollar. Fast mode costs double. In the Claude app, it is included in paid plans that have Fable access, within usage limits.
What is the difference between Fable 5.2 and Mythos 5.2?
They are the same model. Fable 5.2 has extra safeguards and is public. Mythos 5.2 has the cyber and biology capabilities open and is limited to verified organisations.
Can I get Mythos 5.2?
Only if your organisation qualifies through Anthropic's cyber or life sciences verification programmes, or uses Claude Security as a Claude Enterprise customer. Individuals cannot simply sign up for it.
Is Claude Fable 5.2 available in Nigeria?
Yes. Nigeria is on Anthropic's supported countries list, and Anthropic says Fable 5.2 is available on all its platforms. The main practical issues are paying in dollars and keeping API costs under control.
Is Fable 5.2 better than Opus 5.5?
On many of Anthropic's benchmarks, yes. But Anthropic itself recommends Opus 5.5 as the starting point for most coding and knowledge work, and suggests moving up to Fable only for the largest tasks.
Comments (0)
Sign in to post a comment.
- Be the first to comment.


