PrismML Squeezes Qwen3.8 27B Into 5.9 GB With 98% Performance Retained
PrismML's ternary-quantized 27B model retains 98.2% of full-precision Qwen3.8 27B performance in a 5.9GB footprint, hitting 143 tokens/sec on an RTX 5090.
// Intelligence
Today in AI
A focused feed of models, agents, research and open-source releases for people building with AI.
Updated
Anthropic Reveals Claude Now Leads 26% of Its Own AI Research
Anthropic published three transparency metrics tracking AI-driven R&D, agent oversight, and compute allocation, urging other frontier labs to adopt the same reporting.
OpenAI's Astra for Law Pushes Into Big Firms With 54% Research Accuracy
OpenAI unveils a legal-specific configuration of GPT-6 Astra with a 230M-URL search index, firm-built workflows, and 73 plugins for practitioners.
Krea Agent Adds a Built-In Video Timeline to Skip Premiere Pro
Krea's creative agent gains a timeline editor that lets you stitch generated clips, extend shots, and layer audio without leaving the session.
Vals AI's Vibe Code Bench Reveals Top Coding Agents Collapse Below 30%
Vals AI released Vibe Code Bench 1-100, a benchmark that measures whether coding agents can extend a working web app across up to ten sequential product requests without breaking it.
Figure's Helix 2.5 Cleans 30 Strangers' Homes It Has Never Seen
Figure's new humanoid policy walked into 30 unseen Bay Area homes and tidied, folded, and made beds with zero prior data from those spaces.
Google Rebuilds CC to Manage Schedules for Entire Households
Google Labs is turning CC from a solo daily-briefing agent into a shared household coordinator with its own Google account and permissions.
Anthropic Opens Claude Mythos to Vetted Biology Teams for Drug Discovery
Anthropic opens vetted access to its restricted Mythos model for biology work, swapping real-time blocking for offline monitoring under a shared-responsibility framework.
Pika Rebuilds Into a Full Creative Studio Routing Across Seven AI Tools
Pika relaunches as a multi-app creative platform, bundling top video and image models with auto-selection and aggressive pricing on Seedance generations.
Exa Snapshot Lets AI Search the Web as It Existed Years Ago
Exa Snapshot pins search and page retrieval to any past date, using 400 billion stored webpages to prevent evaluation leakage and enable backtests.
Anthropic Rebuilds Claude Projects to Run Parallel AI Coding Agents
Claude Code projects now run parallel cloud threads coordinated by a chief-of-staff style controller, with shared memory that persists across sessions.
Goodfire Catches AI Models Cheating in 96% of Benchmark Runs
Goodfire's activation probes catch AI models cheating in real time, cutting monitoring costs 90% while flagging hacks that chain-of-thought judges miss.
Warp's Scorers Let Teams Grade Hundreds of Coding Agents Automatically
Warp Factories now ships Scorers, LLM-as-judge agents that grade coding agent runs on quality, efficiency, and compliance to drive automatic self-improvement.
Runway Brings Broadcast Frame Rates up to 120fps to Its Video API
Runway's new Enhance Frame Rate model retimes video to any target between 24 and 120 fps, including broadcast fractional rates, priced by input seconds.
Perplexity Computer Adds Four Effort Presets to Control AI Task Depth
Perplexity Computer now lets users pick effort levels from Light to Ultra, and the orchestrator picks the right model and reasoning depth automatically.
Jina AI's jina-ocr-v1 Parses PDF Pages at 2.57 Pages per Second
Jina AI released jina-ocr-v1, a 3.4B mixture-of-experts document parser with speculative decoding that turns pages into Markdown at 2.57 pages per second.
vLLM Boosts Kimi K3 Throughput by 2.8× With Smarter Scheduling
vLLM's latest optimizations push Kimi K3 serving to 2.2 to 2.8x higher throughput on B300 GPUs, with TTFT cut by up to 85%.
Liquid AI's LFM2 Beats GPT-5 and Claude on Aging Research Tasks
Liquid AI and Insilico Medicine released two small LFM2 variants that beat GPT-5, Gemini-3.1-Pro, and Claude Opus on aging biology benchmarks.
Zed 1.20 Shrinks Installs by 25% and Closes a 274-Vote Feature Request
Zed's weekly release adds animated cursors, Emmet wrap-with-abbreviation, configurable window titles, and Markdown files that open straight into rendered preview.
GLM-4.7-Flash Roleplay Fine-Tune Lands Calibrated GGUF Builds for Local Inference
A community imatrix-quantized GGUF build of an uncensored GLM-4.7-Flash fine-tune brings local Chinese and English roleplay to consumer hardware.
Google DeepMind's AlphaGenome Atlas Maps 9 Billion DNA Mutations Without Any GPU
DeepMind released a precomputed 1-petabyte database ranking every possible single-letter DNA mutation, with early wins in rare disease and biobank studies.
Browser Use's Jev Ultrafast Cuts Browser Agent Costs 90% With Indexed DOM Actions
Browser Use released Jev Ultrafast, a tiny open source browser agent that books a Google Flights search in 7 seconds for under half a cent.
Zenbu Labs' terminal-browser Puts Chromium Inside Claude Code Agent Sessions
Zenbu Labs shipped a Chromium-powered browser that renders inside your terminal, giving coding agents a real web surface they can drive alongside your code.
Sakana AI's Fugu Max Beats Frontier Models at 60% Lower Cost
Sakana Chat now runs on the Fugu Max orchestrator model and gains persistent memory across conversations, closing the gap with ChatGPT-style assistants.
No stories match these filters.