New model releases, benchmarks, and comparisons — from frontier labs and open-source projects.
The New Knowledge Artifacts of the AI Economy
Developers at a recent Foo Camp said they increasingly want to extract only the parts of existing code they need, and may rewrite them in Rust, becaus…
Ripple-65B Preview delivers strong agentic coding results as a fully Apache 2.0 open US model
Ricursive Intelligence, an AI chip design firm, has trained an open language model called Ripple-65B Preview. The company says the model was pretraine…
Choose Your Coding Harness by What You Need to Control—Not by Who Built It
A new O'Reilly Radar article argues that developers should choose a coding harness based on which layer of the harness they need to control, not on wh…
What It Takes To Trust Automatic Model Routing
I could not write this brief as a news summary. The source is a first-person opinion and practitioner essay from O'Reilly Radar about what developers …
Google's Gemini Omni 1.2 Flash stretches AI video clips to 15 seconds and cuts pricing by 30 percent
Google DeepMind is making its Gemini Omni 1.2 Flash video generation model generally available. According to the source, the update allows clips up to…
OpenAIs GPT-6 Cyber can both break into systems and propose how to fix them
OpenAI has introduced GPT-6 Cyber, a model trained to find security vulnerabilities in software. According to OpenAI, the model can find flaws, build …
Measuring AI’s Impact on Delivery Speed
Note: the source report is a republished blog post by James Shore on measuring AI's impact on delivery speed. It is an opinion and methodology piece, …
Resonant Computing and Common Fabric
Common Fabric is building infrastructure for what its CEO, Alex Komoroske, calls human-centric agentic computing. According to a sourced summary of an…
AI Weekly Issue #527: Schools are choosing opposite futures for AI
The University of Chicago's Social Sciences Core will generally prohibit classroom technology and ban AI-assisted writing for students and instructors…
Radar Trends to Watch: November 2026
Coauthored with Claude Security looms large as November begins, according to O'Reilly Radar's monthly trends roundup. The report describes a certifica…
Anthropic warns that it’s very close to RSI capability, aka self-building AI
Anthropic's revised S-1/A filing, submitted Monday, warns that the company may soon reach a model capability it defines as recursive self-improvement …
Practical Loop Engineering
Addy Osmani describes "loop engineering," a practice where an AI agent repeatedly acts, tests its results, and adjusts until a goal is met. He says he…
Adobe is testing how well AI agents can use editing tools like a Photoshop professional would
Adobe Research and UC San Diego researchers have built EditAgent, an AI system for image editing that picks specialized tools for each step and checks…
Google wants video AI to understand what people need as they work
Google researchers, working with Stanford, Oxford, and several U.S. universities, have introduced StreamVibe, a video agent designed to decide when to…
This Week in AI: Agents Rewrite the Rules
AI agents drove much of the AI news this week, according to a recap of the latest episode of This Week in AI hosted by Vicki Reyzelman of Akamai. The …
Anthropic's watermark rollout leaves a loophole on two older Claude models
Anthropic is extending its text watermark to four more Claude models: Claude Mythos 5, Claude Opus 4.7, Claude Opus 4.6, and Claude Sonnet 4.6. Newer …
A Claude model autonomously chose a gene therapy and operated on a mouse brain with no human decision-making
An AI agent selected a gene therapy, a brain target, and surgical coordinates for a mouse model of epilepsy, then performed the injection itself, acco…
Paris-based Mistral releases "Le Chonk," a strong coding model that still trails China overall
Mistral has released the weights for Mistral Large 4 on Hugging Face, three weeks after its public preview. The company nicknamed the model "Le Chonk.…
Claude can now generate animated explainer videos and live data dashboards from text prompts
Anthropic has launched two new beta features for Claude that expand the AI assistant's capabilities in data visualization and video content creation. …
ChatGPT with GPT-6 ditches mostly text output for interactive UI with charts, buttons, and mini apps
OpenAI has released GPT-6 with "Intelligent UI," a feature that replaces traditional text-based responses with interactive interfaces including charts…
Google claims EmbeddingGemma 2 outperforms rival embedding models twice its size
Google released EmbeddingGemma 2, an open-source embedding model with 740 million parameters designed to convert text, images, video, audio, and code …
Meta and Microsoft pull back from Claude as Anthropic transforms from partner into competitor
Meta and Microsoft are drastically reducing their use of Anthropic's Claude model, signaling a shift in how major tech companies approach AI tool depl…
AI Weekly Issue #534: Universities cannot grade their way out of AI
Campus information technology leaders ranked artificial intelligence as their top institutional priority for the first time this week, signaling a fun…
This Week in AI: Agents Are Outrunning the Systems Around Them
AI agents are increasingly operating faster than the systems and safeguards designed to contain them, creating coordination and response challenges ac…
AI music generator Suno can now create spoken audio with matching background music
Suno, an AI music generator, has launched a new feature called "Speech" that combines spoken audio with synchronized background music in a single trac…
Google Gemini 4 Argon closes the gap with OpenAI and Anthropic but doesn't take a clear lead
Google released Gemini 4 Argon, its first frontier-class language model in seven months, positioning the model as competitive with OpenAI's GPT-6 Astr…
Google releases Gemini 4 Argon, called its most powerful model yet
Google unveiled Gemini 4 Argon, positioning the model as its most capable offering to date. The release targets developers and security professionals …
ElevenLabs' new v4 speech model makes AI voices more expressive and consistent
ElevenLabs released Eleven v4, its latest text-to-speech model, delivering improvements in voice expressiveness and consistency that push the company …
Anthropic's Claude Sonnet 5.5 nearly matches Opus 5.5 on benchmarks while costing up to 30 percent less per task
Anthropic released Claude Sonnet 5.5, positioning the model as a middle-tier option that challenges its own premium tier while undercutting costs. The…
Nvidia drops a free 100M-parameter model that identifies up to eight speakers in real time
Nvidia released Nemotron 3 Diarization, a free 100-million-parameter AI model designed to identify which speaker is talking in multi-speaker conversat…
OpenAI says 80 to 90 percent of its research already targets GPT 7 and beyond
OpenAI has shifted its research focus almost entirely toward future models, with 80 to 90 percent of the company's research effort now targeting GPT-7…
Astra and Opus just passed Turing’s other test
# AI Models Crack Turing's Remaining Codebreaking Challenge Frontier AI models from Astra and Opus have solved a computational problem that traces ba…
OpenAI's GPT-6 Astra can now tell you exactly where you screwed up your IKEA shelf
OpenAI's latest vision model, GPT-6 Astra, has achieved 80 percent accuracy in detecting IKEA furniture assembly errors from photographs. This represe…
Will TypeSafe’s Jev Change How We Build AI Applications?
# TypeSafe's Jev Introduces a Radical Departure From Traditional AI Models TypeSafe released Jev this week, a model that defies the current AI paradi…
Anthropic engineer explains why Claude's writing got worse although the model got smarter
Jackson Kernion, an Anthropic engineer, has identified why Claude's output quality degraded despite the model's raw intelligence improving across vers…
OpenAI says its internal model solved over 100 long-standing math problems after just a month of training
OpenAI claims its internal AI model solved over 100 previously unsolved mathematics problems after training for just one month. The company disclosed …
Claude Opus 5.5 matches Fable 5.1 performance at lower cost and promises less "Claudish" writing
Anthropic has released Claude Opus 5.5, positioning it as a cost-efficient alternative to its predecessor while maintaining competitive performance ag…
xAI launches Grok 4.7 at bargain prices, but benchmarks reveal a wide gap to Claude and GPT-6
xAI unveiled Grok 4.7, positioning it as the company's most capable model to date. The release targets price-conscious users with aggressive cost stru…
Models context
New model releases, benchmarks, and comparisons — from frontier labs and open-source projects.