Models & Labs
Anthropic Releases Claude Fable & Mythos 5
Matt Wolfehigh confidence
Why it matters
- →Enhances AI transparency and security.
- →Responds to user feedback for improved models.
- →Strengthens Anthropic's position in AI development.
Anthropic has announced the release of Claude Fable & Mythos 5, the latest iteration of its AI models. These updates are designed to improve transparency and address national security concerns. The release follows feedback on previous versions and aims to provide more robust AI solutions.
Read originalThe story around this
Earlier coverage that leads up to this article, and what followed. Lines connect each piece to the closest one after it, converging here.
How we got here
- 1Anthropic Launches Claude Fable 5 and Mythos 5
Anthropic · June 9, 2026 · Same story
- 2Anthropic Releases Claude Fable 5 with Safety Measures
WIRED AI · June 9, 2026 · Same story
- 3Anthropic Launches Fable 5 Model
Sam Witteveen · June 9, 2026 · Same story
- 4Anthropic Releases Claude Fable 5 and Mythos 5
Skill Leap AI · June 9, 2026 · Same story
- 5Anthropic Releases Claude Fable 5 to Public
The Rundown AI · June 10, 2026 · Same story
- 6Anthropic Releases Claude Fable 5
AI Explained · June 10, 2026 · Same story
- 7Anthropic Releases Fable 5 AI Model
Matt Wolfe · June 11, 2026 · Same story
- 8Anthropic Releases Claude Fable Model
Fireship · June 11, 2026 · Same story
What happened next
- 9Anthropic Revives Fable 5 and Launches Sonnet 5
The AI Advantage · July 3, 2026 · Same story
- 10Anthropic launches Claude Fable 5.1 with cost savings
The Verge AI · September 1, 2026 · Same story
- 11Anthropic launches Claude Fable 5.1 model
The Rundown AI · September 2, 2026 · Same story
- 12Anthropic Introduces Claude Fable and Mythos 5.1
Matt Wolfe · September 3, 2026 · Same story
- 13Anthropic Releases Fable 5.1 and Mythos Models
AI Explained · September 4, 2026 · Same story
Follow this story
Open the full story →Microsoft Limits Claude Fable Use Over Data Concerns
17 developments
- Jun 10 · The Verge AI
Microsoft Limits Claude Fable Use Over Data Concerns - Jun 10 · The Verge AI
Anthropic's Claude Fable 5 Limits Biology Queries↳ Claude Fable 5 blocks basic biology queries like mitochondria due to strict bioweapons safeguards despite strong cybersecurity performance. - Jun 11 · WIRED AI
Anthropic Reverses Controversial Claude Fable 5 Policy - Jun 11 · The Verge AI
Anthropic Apologizes for Hidden Claude Fable Guardrails - Jun 12 · The AI Daily Brief
Anthropic's Fable-5 Launch Sparks Controversy↳ Fable-5 launch sparked backlash over opaque safety filters and silent model degradation. - Jun 12 · Matt Wolfe
Anthropic Releases Claude Fable & Mythos 5 (This article)↳ The new model is named Claude Fable & Mythos 5 and enhances AI capabilities. - Jun 12 · The AI Advantage
Claude Fable 5 Capabilities Explored - Jun 13 · WIRED AI
Anthropic Takes AI Models Offline Due to US Order↳ US government ordered Anthropic to take Claude Fable 5 and Mythos 5 offline citing national security concerns. - Jun 13 · TechCrunch AI
U.S. Government Halts Anthropic's AI Models Over Security Concerns - Jun 13 · Duncan Rogoff
US Government Shuts Down Anthropic's Claude Fable 5 - Jun 13 · Cole Medin
US Government Halts Anthropic's Claude Fable 5 - Jun 13 · The AI Daily Brief
US Government Orders Shutdown of Anthropic's Fable 5 - Jun 14 · TechCrunch AI
Anthropic Suspends AI Model Access Amid U.S. Directive - Jun 14 · AI Explained
Anthropic Admits Ban Will Cause Difficulties↳ Anthropic admits the ban on Claude Fable 5 will cause significant difficulties. - Jun 17 · The AI Daily Brief
Anthropic Shuts Down Mythos and Fable Due to Jailbreak Risks↳ Anthropic cites jailbreak vulnerabilities and cybersecurity weaknesses as the specific reasons for shutting down Mythos and Fable. - Jun 19 · Lev Selector
Claude Fable 5 Withdrawn Amid Negotiations↳ Anthropic is negotiating with the administration to regain access to Claude Fable 5. - Jun 24 · Matt Wolfe
Anthropic Suspends Foreign Access to Fable 5
More in Models & Labs
Models & Labsother
llama.cpp b12340 adds CUDA 13 and ROCm 10.0
This release quietly closes the hardware gap for local inference by adding default support for CUDA 13 and ROCm 10.0 across Linux and Windows. Users with newer NVIDIA GPUs or AMD accelerators no longer need to compile from source to get performance, while Apple Silicon builds now ship KleidiAI disabled by default, shifting the focus back to standard ARM optimizations. The inclusion of Snapdragon support for Linux and Android marks a significant step toward on-device AI beyond the x86/ARM desktop ecosystem. It is less about new features and more about ensuring llama.cpp remains the universal runtime for every available accelerator.
llama.cpp Releases
Models & Labscoding
llama.cpp b12342 adds few-row MMQ for speculative decoding
Speculative decoding just got a significant speed boost on NVIDIA GPUs. This release introduces a specialized kernel for 'few-row' matrix multiplications, targeting the small batch sizes typical of verification passes in speculative decoding workflows. By optimizing these specific operations for quantized formats like Q8_0 and IQ4 variants, llama.cpp reduces latency where it matters most during token generation. The change is CUDA-only for now but leverages hardware capabilities from Ampere onwards, making local inference faster without requiring new model architectures or complex configuration changes.
llama.cpp Releases
© TechCrunch AIModels & Labsmodels
Google deploys Gemini 4 Hydrogen for faster Search
Google is quietly shifting its search infrastructure by replacing older fast models with Gemini 4 Hydrogen, a tier designed specifically for speed and cost-efficiency rather than raw reasoning power. This model completes tasks 2.5 times faster than its predecessor while beating the more powerful Argon on specific intelligence benchmarks, proving that specialized efficiency can outperform brute force in high-volume workflows. By integrating this into Search's fan-out technique, Google now breaks down complex queries into subqueries with greater precision and speed, delivering richer visual responses without the latency of heavier models. This marks a strategic pivot toward agentic execution layers within consumer search, making AI Mode feel snappier for everyday multi-part questions.
TechCrunch AI