[AINews] Claude Opus 5: Fable-level performance at Opus price (half Fable)
Anthropic releases Claude Opus 5 matching Fable performance at half the cost, demonstrating efficiency gains in model distillation.
Top Stories
All Stories →Introducing Claude Opus 5
Anthropic releases Claude Opus 5, matching Fable 5 frontier performance at half the cost, now leading Artificial Analysis leaderboard.
Ruff v0.16.0
Ruff v0.16.0 enables 413 default linting rules (up from 59), breaking existing CI pipelines and catching syntax/runtime errors previously uncaught.
Quoting Boris Cherny
Claude Opus 5 achieves lowest prompt injection vulnerability rate across evals and red team testing, per Anthropic's system card.
Introducing Claude Opus 5
Anthropic releases Claude Opus 5 with improvements in agent execution, coding, and professional tasks.
[AINews] Black Forest Labs FLUX 3 - Multimodal Flow Models that beat Seedance 2.0, Gemini Omni and Grok Imagine, and FLUX-mimic video-action robotics model
Black Forest Labs releases FLUX 3 multimodal model with reported improvements over Gemini 2.0, Grok Imagine, and includes video-action robotics variant.
Kimi K2.6 Released (huggingface)
Kimi K2.6 released on Hugging Face; availability announcement for open-weights download.
Deep Dives
Librarians are hosting viral ‘Avoiding AI’ workshops for people who are fed up with Big Tech
At libraries around the country, "Avoiding AI" workshops have elicited unprecedented demand.
·
One fallen power line exposed a growing AI data center problem. Here’s how to fix it.
A close call in Northern Virginia revealed just how poorly data centers respond to grid disruptions. Here's how to fix the problem.
·
I tried out OpenAI’s new AI keypad — which will be fun for some coders and slightly mystifying to everyone else
OpenAI's fancy new AI keypad will be a lot of fun for some, while many others are probably not going to touch it.
·
Prentis, new AI lab co-founded by Reid Hoffman, Marc Pincus in talks to raise $100M
The neolab is betting that automating routine computer tasks will soon outpace coding as AI's biggest use case.
·
Canadian legislator reads out apparent LLM response in floor speech
"Here’s a more natural, flowing version of that section..."
·
Anthropic's Opus 5 is about token efficiency, not a capability leap
Models are improving quickly, but the cheaper options are often good enough.
·
Also on the wire
- You can’t ignore Google Zero anymoreThe Verge AI
- 2026.30: The Copium WarsStratechery
- Anthropic launches Opus 5TechCrunch AI
- Anthropic releases Opus 5 with ‘close’ to Fable 5’s capabilitiesThe Verge AI
- Meta is making its AI chatbot more like an assistantThe Verge AI
- ModelExpress: Distributing Model Artifacts at the Speed of LightNVIDIA Dev Blog
- As US weighs response to Chinese AI, industry urges against broad open-weight restrictionsTechCrunch AI
- Bluesky’s AI assistant Attie expands into an open social research toolTechCrunch AI
- Midjourney acquired the astrology app Co-StarTechCrunch AI
- The tech-broification of American science has officially begunThe Verge AI
- ‘AI communism’, rogue models, and the why Kimi K3 spooked Wall StreetTechCrunch AI
- AI firms want more data centers; Trump's EPA may give neighbors less sayArs Technica AI
- OpenAI’s new voice mode makes it to the ChatGPT desktop appTechCrunch AI
- How AI guardrails are impeding the work of offensive cybersecurity researchersTechCrunch AI
- The first known runaway AI agent - or a very bad marketing stunt?Simon Willison
- Alexa Plus is getting an AI update to handle more complicated instructionsThe Verge AI
- AMD takes on Nvidia with its Helios AI rack scale systemTechCrunch AI
- AI Kill Switch Act would let Trump admin order shutdown of rogue AI systemsArs Technica AI
- Anthropic updates Claude voice mode with more capable modelsTechCrunch AI
- Claude’s voice mode is now available for Opus and SonnetThe Verge AI
- AegisAI, founded by former Google security execs, lands $36M to stop AI-driven spear phishingTechCrunch AI
- Patreon is laying off 20 percent of workersThe Verge AI
- 3D-Aware VLMs with Implicit and Explicit GeometriesarXiv (cs.AI/CL/LG)
- Expanding Flow MapsarXiv (cs.AI/CL/LG)
- GraphVid: Interactive Graph-Controllable Video GenerationarXiv (cs.AI/CL/LG)
- Barzilai-Borwein Fails Superlinear Convergence on an Open Set of Quadratics for Every Dimension $n\geq 4$arXiv (cs.AI/CL/LG)
- Synthetic data generation framework for quality control automation in gravure printingarXiv (cs.AI/CL/LG)
- Surprisal Theory is Tautological (without Rational Grounding)arXiv (cs.AI/CL/LG)
- Beyond Sufficiency: Time Series Explanation with Counterfactual NecessityarXiv (cs.AI/CL/LG)
- MedGame: Storytelling Gamification Empowered by Large Language Models for Medical EducationarXiv (cs.AI/CL/LG)
- Graph Learning on Ensembles of Cyclic Peptides: An Investigation of Molecular Ensemble ModelingarXiv (cs.AI/CL/LG)
- Unsupervised Consensus-Based Anomaly Detection for Spatiotemporal Malaria Incidence in GhanaarXiv (cs.AI/CL/LG)
- Beyond Sycophancy: Structured Resistance and Compliance in LLM Moral ReasoningarXiv (cs.AI/CL/LG)
- OpenForgeRL: Train Harness-native Agents in Any EnvironmentarXiv (cs.AI/CL/LG)
- Visual Contrastive Self-DistillationarXiv (cs.AI/CL/LG)
- MIRROR: Learning from the Other View for Multi-Modal ReasoningarXiv (cs.AI/CL/LG)
- X$^3$-OPD: Distilling Reasoning into Large Audio-Language Models via On-Policy AlignmentarXiv (cs.AI/CL/LG)
- Neural solutions of coupled ghost and gluon Dyson--Schwinger equations in Landau gaugearXiv (cs.AI/CL/LG)
- The Boundaries of Automation: A Theory of Persistent Human ParticipationarXiv (cs.AI/CL/LG)
- Zero-Flow Two-Sample TestsarXiv (cs.AI/CL/LG)
- Windowed-MTP: Removing the Full-Context Draft-KV Tax at Million-Token ContextarXiv (cs.AI/CL/LG)
- From Resource Flow to Executable Tests: Petri-Net-Guided LLM Test Generation for Concurrent Stateful Rust APIsarXiv (cs.AI/CL/LG)
- ElasticTTT: Prior-Preserving Test-Time Tuning for Video EditingarXiv (cs.AI/CL/LG)
- Runway launches AI model router as generative media gets crowdedTechCrunch AI