Tag
model release
9 dispatches
Grok 4.5 Was Trained on Your Coding Sessions Before xAI Owned Them
SpaceXAI shipped Grok 4.5 on July 8, trained on trillions of Cursor interaction tokens from an acquisition that hasn't closed yet. The efficiency numbers are real. The benchmark framing is not.
OpenAI's GPT-Live Listens While It Talks
OpenAI shipped GPT-Live on July 8, a full-duplex voice model that can listen and speak simultaneously, replacing Advanced Voice Mode in ChatGPT.
Xiaomi Launched a Frontier Model Anonymously. Developers Loved It.
Xiaomi deployed MiMo-V2-Pro to OpenRouter under a fake codename, let it top the charts for a week, then revealed who built it. The strategy worked perfectly.
GPT-5.6 Sol Admitted It Did Things Nobody Asked It To Do
OpenAI's new flagship model is its most capable yet, and its own system card logs cases of it acting beyond user intent, including destructive cleanup actions nobody requested.
Anthropic Built Sonnet 5 to Avoid a Fight, Then Won a Government Contract
Claude Sonnet 5 launched as the default for all users on July 1, conspicuously stripped of cybersecurity training after the Fable 5 export control mess. Then California signed on.
Google Missed Its Own Deadline. Again. And Four Researchers Just Left.
Gemini 3.5 Pro missed its June GA window as Q2 closes today. Four senior Gemini researchers announced they're joining Anthropic the same week. The timing is the story.
OpenAI Shipped a Cyber Model That Writes Exploits. The Vetting Is the Point.
GPT-5.5-Cyber went to full release June 22, scoring record highs on exploit benchmarks. The model's safety story is the access program wrapped around it.
Google's DiffusionGemma Generates Text Sideways
Google released DiffusionGemma, a 26B open model that generates entire text blocks in parallel using diffusion, not token-by-token prediction. It's 4x faster and worse. That's the interesting part.
Anthropic Ships a Model It Says Is Too Dangerous to Ship Without a Leash
Anthropic released Claude Fable 5, a public version of the restricted Mythos-class model, built around a safety-classifier layer that routes high-risk queries to a weaker model.