Claude Fable 5 is here: Anthropic's first public Mythos-class model, with a safety wall built in
Anthropic launches Claude Fable 5 with a 1M-token context window, $10/$50 pricing, and a safety-classifier fallback — plus a restricted Mythos 5 for Project Glasswing partners.
Anthropic has shipped Claude Fable 5, its most powerful model to date and the first Mythos-class model available to the general public. At the same time, it quietly launched Claude Mythos 5 to a restricted group of cybersecurity partners. The two releases are, in a technical sense, the same underlying model. What separates them is a set of safety classifiers, and Anthropic’s confidence in who is on the other side of the API call.
Update, 13 June 2026: Anthropic later suspended access to Fable 5 and Mythos 5 after a US government export-control directive. The launch details below are useful background, but the current status is covered in the follow-up article on the suspension.
What is a Mythos-class model, and why does the name matter?
Up until now, Opus was the top tier in Anthropic’s model family. Mythos sits above it. The naming is deliberate: “fable,” from the Latin fabula, is the Latin cousin of the Greek mythos. Fable 5 is the version made safe for general use. Mythos 5 is the same model with some of those safeguards lifted, available only to vetted partners.
This is a new structural approach for frontier AI. Rather than releasing one model and hoping misuse stays low, Anthropic has split a single capability level into two distinct products separated by a safety layer.
What Fable 5 can actually do
The benchmark numbers are significant. On SWE-bench Verified, the standard software engineering evaluation, Fable 5 scores 95.0%. Claude Opus 4.8 sits at 69.2%. GPT-5.5 scores 58.6%. On the harder SWE-bench Pro, Fable 5 reaches 80.3% against Opus 4.8’s 13.4%.
To put that in real terms: Stripe reported during early testing that Fable 5 completed a codebase-wide migration across a 50-million-line Ruby codebase in a single day. The same task, done by hand, would have taken a whole engineering team more than two months.
Vision capabilities have also taken a meaningful step forward. The model can rebuild a web app’s source code from screenshots alone. In a more unusual proof of capability, Anthropic says Fable 5 completed Pokémon FireRed with a minimal agentic harness. That is a long-horizon reasoning task with hundreds of sequential decisions. It is not a party trick.
Context window and pricing
Fable 5 ships with a 1 million token context window as standard, with up to 128,000 output tokens per request. Critically, there is no long-context surcharge. A 900,000-token request is billed at the same per-token rate as a 9,000-token one.
Pricing is $10 per million input tokens and $50 per million output tokens. Batch pricing drops to $5 and $25 respectively. Anthropic notes this is less than half the price of Claude Mythos Preview. The model IDs are claude-fable-5 and claude-mythos-5.
Adaptive thinking is always on. There is no way to disable it. The effort parameter controls thinking depth, but the raw chain of thought is never returned. This is a deliberate design choice, not a limitation of the API.
The safety classifier fallback
Here is the part that will matter most to developers building on top of the model. Fable 5 does not simply refuse high-risk queries. It routes them to Claude Opus 4.8 instead.
When that happens, the Messages API returns a stop_reason: "refusal" with an HTTP 200, not an error. The response includes a stop_details.category field that tells you which classifier fired: “cyber”, “bio”, “reasoning_extraction”, or null. Anthropic says this fallback triggers in fewer than 5% of sessions on average, and that the classifiers are tuned conservatively.
Billing follows the routing. If your prompt is routed to Opus 4.8, you pay Opus prices for the tokens processed there. If a request is blocked mid-conversation, the initial tokens are charged at Fable rates and the subsequent ones at Opus rates.
Where it launched, before access was suspended
At launch, Fable 5 was live on the Claude API, Claude Platform on AWS, Amazon Bedrock (US East N. Virginia and Europe Stockholm regions), Vertex AI, and Microsoft Foundry. Access was later suspended after the export-control directive noted above.
On subscriptions, the original timeline was tight. Pro, Max, Team, and seat-based Enterprise plans were due to include Fable 5 at no extra cost through 22 June. From 23 June, accessing the model was due to require usage credits. That plan was overtaken by the suspension, so it should be read as launch context rather than current availability guidance.
One more change worth flagging: Anthropic is now requiring 30-day data retention on all Fable 5 and Mythos 5 traffic, including for enterprises that previously held zero-retention agreements. Anthropic says the data will not be used for training and will only be used to detect patterns of misuse across sessions. That is a reasonable justification, but if your organisation had zero-retention as a contractual requirement, this is worth checking with your legal team.
Claude Mythos 5 and Project Glasswing
Running alongside the public launch, Anthropic also upgraded its Project Glasswing partners to Claude Mythos 5. Project Glasswing began in April, giving Mythos Preview to around 150 organisations including financial institutions, healthcare networks, and critical software infrastructure providers. The restricted programme operates in collaboration with the US government.
Mythos 5 has safeguards lifted in certain areas, making it suitable for defensive cybersecurity work where the full capability of the model is needed. A separate biology programme will give selected life sciences researchers access to Fable 5 with biology and chemistry safeguards removed while cyber safeguards remain in place.
What this means for you
If you are a developer, the immediate priority is understanding the refusal routing before it surprises you in production. A 200 response with stop_reason: "refusal" is not an error, but it will behave differently to a normal completion. Build for it explicitly.
If you are on a subscription plan and want continued access to Fable 5 after 22 June, start calculating your likely usage now. Switching to usage credits mid-project is easier if you have planned for it.
If you are in enterprise software, the coding benchmark numbers are hard to ignore. The gap between Fable 5 and its nearest competitors on tasks like large-scale codebase migration is wide enough to have a material effect on engineering throughput.
And if your organisation had zero-retention agreements with Anthropic on Bedrock or the API, the new 30-day retention requirement needs a conversation with your compliance team before you migrate to Fable 5.
Fable 5 was available at launch, but access has since been suspended while Anthropic works through the government directive. Treat the subscription access window and pricing details above as launch context, not current availability.
Updates to this story
7 July 2026: Anthropic extends Claude Fable 5 subscription inclusion to July 12 after user backlash
Hours before Fable 5 was set to exit standard paid plans on July 8 (12 AM PT), Anthropic extended promotional access to July 12, 2026, at 11:59:59 PM PT following user backlash. No action is required to receive the extension; it applies automatically to eligible accounts.
The extension applies to Pro, Max, Team, and premium Enterprise seats where the feature has been enabled. During this window, subscribers can use up to 50% of their weekly usage limits on Fable 5 at no extra charge, since the model draws from the standard weekly allocation at a faster rate than other Claude models. Legacy seat-based Enterprise standard seats, usage-based Enterprise plans, and API access are not covered.
This supersedes the original story’s framing around the July 8 cutoff, but it does not change the underlying access model. After July 12, Fable 5 exits the weekly subscription bucket entirely, and continued access requires prepaid usage credits at $10 per million input tokens and $50 per million output tokens. If credits are not enabled, access stops with no grace period.
Anthropic has said it wants to restore Fable 5 as a standard subscription benefit when compute capacity allows, but has given no timeline. July 12 remains the date to plan around. To enable usage credits after the window closes, go to claude.ai > Settings > Usage on the web app.
Sources: Anthropic release notes, Android Authority
1 July 2026: Claude Sonnet 5 launches as the new default model across all plans
Anthropic launched Claude Sonnet 5 on June 30, 2026, making it the new default model for Free and Pro users and replacing Sonnet 4.6. Max, Team, and Enterprise subscribers can also select it, and developers can access it via the Claude API, Amazon Bedrock, Google Cloud Vertex AI, and Microsoft Foundry.
This reshapes the context this article was written in. Claude Fable 5 and Mythos 5 remain suspended worldwide following the June 12 Commerce Department order. Sonnet 5 and Opus 4.8 are now the effective ceiling for most developers and users.
Sonnet 5 carries a 1M-token context window and 128k max output, with adaptive thinking enabled by default. Introductory API pricing is $2/$10 per MTok through August 31, rising to $3/$15 after that date. Note that an updated tokenizer means the same input can produce up to 1.35x more tokens than with Sonnet 4.6, though Anthropic designed the introductory pricing to keep the transition roughly cost-neutral. Agentic coding performance reaches 63.2%, up from Sonnet 4.6’s 58.1%.
Anthropic describes Sonnet 5 as a drop-in replacement for Sonnet 4.6 at the same API endpoint. Cyber safeguards are enabled by default, though they are less restrictive than those applied to Fable 5. The August 31 pricing cliff is the most immediate date for API developers to track.
30 June 2026: Claude Sonnet 5 launches as new default model with 1M context and $2/$10 intro pricing
Anthropic has launched Claude Sonnet 5 as its new default model across Free, Pro, Max, Team, and Enterprise plans, plus the API, Claude Code, Amazon Bedrock, and Google Cloud Vertex AI. If you use Claude today, you are almost certainly already talking to Sonnet 5.
This matters for the Fable 5 story because Sonnet 5 now fills the practical, widely-available slot in the lineup that Fable 5 does not. Where Fable 5 targets restricted partner use cases with a built-in safety-classifier fallback, Sonnet 5 brings near-Opus-4.8 agentic performance, planning, and autonomous tool use down to mid-tier pricing for everyday developer work.
Introductory pricing is $2/$10 per million input/output tokens through August 31, 2026, after which standard pricing rises to $3/$15. The more important detail for anyone migrating from Sonnet 4.6: Sonnet 5 uses a newer tokenizer that produces roughly 30% more tokens for the same text. Anthropic set intro pricing to be approximately cost-neutral against Sonnet 4.6, but once standard pricing takes effect, effective costs for most workloads will be meaningfully higher. Audit your token budgets before switching production traffic.
Opus 4.8 remains the stronger option for pure agentic coding tasks, trailing Sonnet 5 by roughly 6 points on SWE-bench Pro.
23 June 2026: Claude Fable 5 exits the free subscription window and credit-only billing begins
As of today, Claude Fable 5 is no longer included in Pro, Max, Team, or seat-based Enterprise subscription plans. The free window that ran from launch through June 22 has closed. Continuing to use Fable 5 now draws from usage credits billed at the same per-token API rates covered in the original post: $10 per million input tokens and $50 per million output tokens, with the existing 90% discount for prompt caching. Every token counts against that balance, including Claude Code sessions, Research mode, and project file content.
This supersedes the original post’s framing of Fable 5 as a subscription-included model. The pricing figures remain accurate; what has changed is how you access those tokens.
If you hold no credits, Fable 5 does not disappear. Claude falls back to your plan’s standard models, including Opus 4.8, and your subscription continues as normal. Anthropic’s stated reason is capacity: demand is difficult to predict, and the company says it intends to restore Fable 5 as a standard subscription feature “as quickly as we can,” with advance notice before any change.
Mythos 5 and Project Glasswing access are unaffected. For teams watching costs, routing routine workloads back to Opus 4.8 and setting spending caps in the console are the two most immediate steps worth taking.
22 June 2026: Claude Fable 5 free subscription access ends June 23 — usage credits required from June 23 onward
Two developments have changed the picture since launch.
The model is currently offline. On June 12, the US government issued an export control directive suspending all access to Fable 5 and Mythos 5, citing national security grounds and a reported jailbreak technique. Anthropic disabled both models globally because it cannot verify user nationality in real time, making this the first time US export controls have been applied directly to an AI model rather than to underlying hardware. Anthropic disputes the action publicly, arguing the cited jailbreak is narrow and replicable using other publicly available models. As of June 22, no restoration timeline has been given; an administration official indicated the suspension could last several weeks. All other Anthropic models, including Opus 4.8, Sonnet 4.6, and Haiku 4.5, remain fully available.
The June 23 billing change still stands if the model returns. Per the original launch terms, Fable 5 has been included at no extra cost on Pro, Max, Team, and seat-based Enterprise plans through June 22. Starting June 23, continued use on those plans requires usage credits. API pricing ($10/$50 per million tokens), consumption-based Enterprise, and cloud platform access via Bedrock, Vertex, and Foundry are not affected. Note that Fable 5 counts roughly double the usage of Opus 4.8 against subscription limits.
Monitor Anthropic’s release notes for restoration updates.