Tag: Large language models

Claude Fable 5: Anthropic unlocks Mythos-class AI for everyone

Anthropic releases Claude Fable 5 — the first Mythos-class model in general availability. Same core as the previously restricted Mythos Preview, wrapped in new classifiers that block high-risk requests. Priced at $10/$50 per 1M tokens.

Pentagon Signs AI Deals with NVIDIA, Microsoft, and AWS for Classified Military Networks

The U.S. Department of Defense has signed AI deployment agreements with NVIDIA, Microsoft, AWS, and Reflection AI for its IL6/IL7 classified networks, accelerating its multi-vendor strategy in the wake of the Anthropic legal dispute.

AI Outdiagnoses Doctors in Harvard Emergency Room Study

OpenAI models o1 and 4o outperformed internal medicine physicians in diagnosing emergency department patients in a controlled study published in Science by Harvard Medical School and Beth Israel researchers. Model o1 achieved 67% accuracy at initial triage versus 55% and 50% for physicians.

xAI Grok 4.3: Always-On Reasoning, New Pricing, New Limitations

xAI has released Grok 4.3 — a model with an always-on reasoning architecture, 1M token context, and prices cut by more than 50%. The model targets enterprise workloads, but independent benchmarks flag weaknesses in coding and math. It comes alongside Custom Voices — voice cloning from a 2-minute sample.

Anthropic Launches Claude Security for Enterprises in Public Beta

Anthropic has released Claude Security in public beta for Enterprise customers. The tool scans codebases for vulnerabilities and generates patches, running on Claude Opus 4.7. The release precedes the broader rollout of the controversial Mythos cybersecurity model.

Musk Admits Under Oath: xAI Used OpenAI Models to Train Grok

Testifying in a California federal court, Musk confirmed xAI used distillation on OpenAI models to train Grok — while suing OpenAI for abandoning its founding principles.

Mixture of Experts (MoE) — architektura, która pozwala budować coraz większe modele AI bez proporcjonalnego wzrostu kosztów obliczeniowych

Zhipu AI uderza w OpenAI. Model GLM-5.1 pracuje autonomicznie przez 8 godzin

Google otwiera Gemma 4: Nowa generacja modeli AI stawia na lokalne działanie i agentowość

EverMind przełamuje barierę 100 milionów tokenów. Nowa architektura MSA redefiniuje pamięć długotrwałą w LLM

Grounding with Google Maps: Jak Google łączy sztuczną inteligencję z danymi przestrzennymi