Anthropic has released Claude Fable 5 as its first publicly available Mythos-class model, which it describes as exceeding any prior generally available model on nearly all benchmarks. The release includes conservative safeguards that redirect some queries to Claude Opus 4.8, while a less-restricted Mythos 5 version stays limited to select cyberdefenders.
Anthropic adds that the longer and more complex the task, the larger Fable 5’s lead over its other models.
Mythos 5 is only available to a small group of cyberdefenders and infrastructure providers.
Tap a lens to see what this story means for you.
Liked this? The Brief brings you the whole day in tech, verified, every morning.
Two minutes, free forever. What's in The Brief →
See what’s happening right now
The Feed runs all day — short, verified briefs the moment they break.
Open the FeedFollow @thecircuitry_
Every story we publish, as it happens. No noise between.
Reader-supported
The Circuitry is a passion project I've always wanted to build, and I love the work behind it.
Running it costs real money. APIs, hosting, time. To keep improving the site and growing this into something useful for everyone, those costs have to be covered.
Any contribution is appreciated. If not, no pressure. Thanks for reading.
Anthropic launched Fable 5.1 and Mythos 5.1 on September 1, 2026, claiming stronger performance at 25 percent lower typical cost and up to 45 percent savings on agentic work. The release also brings refined safeguards, enterprise data-retention improvements rolling out this fall, and expanded but still limited use for software vulnerability identification.
Anthropic released Claude Haiku 5.5, its cheapest and fastest small model, at $0.10 per million input tokens and $0.50 per million output tokens for prompts up to 100,000 tokens. Anthropic says it costs around 75% less to run than Haiku 4.5 on average and scores far higher on its benchmarks.
Anthropic's Claude service is down worldwide due to authentication system failures affecting multiple products. More than 1,500 reports appeared on Downdetector within minutes on September 29.
Anthropic introduced Claude Sonnet 5.5 at unchanged mid-tier rates of $2 per million input tokens and $10 per million output tokens. The model exceeded Opus 5.5 on Terminal-Bench 4.0 yet consumed more tokens at peak effort settings.
Anthropic launched Claude Opus 5.5 on September 22, 2026 as the first model in its Claude 5.5 family, claiming it matches Claude Fable 5.1 performance while running 40% cheaper and 30% faster than Opus 5. The release highlights continued pressure on AI developers to improve efficiency without sacrificing quality.
