OpenAI published ten mathematical and theoretical computer science results produced by an internal variant of its frontier model, Astra, spanning high-dimensional geometry, coding theory, group theory…
Mistral has opened a public preview of Mistral Large 4 (ML4), a one-trillion-parameter open-weight model it trained on 4,000 Nvidia Grace Blackwell GPUs over two months in its own European data center…
A research note from FUNDA estimates that AI computer-use agents could automate work equivalent to roughly 180 million full-time jobs, about 18% of global white-collar employment, representing a long-…
Kairos 1, announced by Kevin Bang He, is described as the first model built to simulate individual human behavior and outperforms Persimmon, Astra, Fable, and Gemini across 13 simulation benchmarks, a…
DeepSeek V4.1 Flash failed to recreate a headshot by operating a web-based replica of Microsoft Paint in a test run by Matthew Berman, producing a portrait Berman called "actually not bad" but "very a…
OpenAI announced GPT-6.1 Sol at DevDay alongside Dots and Spaces, an efficiency-focused model that OpenAI says approaches Astra's performance on selected evaluations while costing roughly a fifth of A…
OpenAI released GPT-6.1 Sol, a model priced at $2 per million input tokens and $10 per million output tokens, which byteiota reports is 5x cheaper than Astra while delivering near-Astra performance. T…
A Graphite study of frontier-model prose found 13,000 phrases at least twice as common in AI-generated text as in human writing, with Claude Opus 5.5 using "this matters" 116 times more often than hum…
A developer's experiment on a durable TypeScript and SQLite reminder queue found that both the Astra solo and Astra + Luna implementations passed their original acceptance suites while still containin…
Fin's 2x Team published a set of context-management practices for Claude Code sessions, arguing that keeping starting context small and pruning stale memory directly lowers token costs and improves ou…
Researchers introduced modal kinetic typography, a method that animates a vector glyph by building motion from the glyph's natural vibration modes, computed via a finite-element eigenproblem assembled…
Halv cut recorded model cost by 57.1% versus vanilla Codex across 14 repository tasks, spending $144.67 against $337.50 while both arms passed 25 of 42 runs, according to Halv's own interim benchmark …
Yolanda and Spencer released a free CLI token compression tool for Codex that cuts input tokens by 29.6% and saves up to 30% on costs, after they burned $700/day per person on API usage. The tool uses…
OpenAI announced several new agent capabilities at DevDay 2026, including always-on "Dots" agents, cloud-based Codex, the GPT-6.1 Sol model, and computer use in its Agents API. A developer outlined a …
OpenAI unveiled "dots" — always-on AI agents that run in their own cloud-based virtual computers — at its developer conference on Tuesday, initially available to Pro, Business Premium and Enterprise u…
OpenAI CEO Sam Altman outlined a three-part plan for the company at its developer conference in San Francisco on Tuesday: provide the best AI models, make them the easiest for developers to build on, …
OpenAI CEO Sam Altman debuted the company's Dots AI agent platform on Tuesday, powered by OpenAI's Astra model, positioning it for the enterprise market with capabilities including triaging Slack noti…
OpenAI restated at its September 29th DevDay keynote that it reached its "automated research intern" milestone by September, a claim first published in its September 6th research report, with the next…
OpenAI announced Dots, a line of always-on AI agents that run continuously in the background and can be reached via text, phone calls, and email, during a livestream. Dots can buy things on a user's b…
OpenAI released GPT-6.1 Sol, a model the company says delivers near-Astra performance at a lower cost for complex coding, computer use, and professional work, priced at $2.00 per 1M input tokens and $…