Asana cut model costs 76x and increased execution speed 5x in browser-agent tests using GPT-6.1 Sol, according to OpenAI. The cost and latency reduction moves browser-based agents from expensive async…
Anthropic's Claude Opus 5.5, shipped September 22, is 40% cheaper than Opus 5, 30% faster, and beats GPT-6 Astra on benchmarks at one-third the cost, but four API changes throw 400 errors when teams f…
KingBench 4, an independent benchmark scoring five AI coding models on six difficult software projects out of 10, found Anthropic's Opus 5.5 leading the hardest tasks with wins on the emulator (3.7/10…
Nace.AI launched Nace Document Intelligence (NDI), a document parsing layer for AI agents, claiming its Nace DI High setting ties the best measured Parse Index score of 0.820 at $6.50 per 1,000 pages …
Hex used 1,800 eval cases and an automated hill-climbing loop to cut the wrong-edit rate of its Quick Edits feature from 21% to 3%, the company reported in a blog post. The Quick Edits feature uses a …
Particle theorist Nathaniel Craig, on a one-year visiting position at OpenAI, posted arXiv preprint 2610.09093, "A gauge-invariant measure for lattice chiral fermions," a 123-page paper extending Mart…
NVIDIA detailed how developers are using frontier AI models, including GPT-6 Astra and Claude Fable 5, to direct agents that assemble simulation applications on top of NVIDIA Omniverse libraries such …
NVIDIA published a five-step workflow for converting CAD assets into SimReady OpenUSD assets for robotics simulation, using an ABB Robotics YuMi robot as the worked example and GPT-6 Astra as one fron…
OpenAI publicly released 722 mathematical manuscripts organized into 372 related result families on October 6, 2026, produced by an internal frontier model from an evaluation involving roughly 4,000 m…
Epoch AI's new Automation Reports, introduced October 8, 2026, found that Anthropic's Claude Fable 5.1 and OpenAI's GPT-6 Astra tied for the top score at approximately 65% on 11 tasks drawn from Epoch…
Meta and Sierra jointly announced the Personal Agent Protocol on October 6, 2026, an open standard built on OAuth that lets consumers set permissions for personal AI agents and businesses define which…
Claude Opus 5.5 ranked first overall among 181 AI model configurations tested across 9,050 generated drafts on 10 writing tasks, but GLM-5.3 Flash scored less than 3% lower at roughly 190 times lower …
A developer's cost analysis finds that running 1,000 cached agentic coding tasks costs $192 on OpenAI's GPT-6.1 Sol versus $1,020 on GPT-6 Astra, a 5.3x gap driven by Sol's $2/$10 per-million input/ou…
OpenAI announced Dots at DevDay on 29 September 2026 in San Francisco, describing them as "always-on agents powered by GPT-6 Astra" that run on a persistent cloud computer with a browser, long-term me…
Three fired OpenAI employees — Tomek Korbak, Mikita Balesni, and Jasmine Wang — wrote a letter to OpenAI's board and safety committees urging that chain-of-thought monitoring be preserved, according t…
A rebuilt buy-side analyst workflow pairs ChatGPT Plus running GPT-6 Astra with Instinct, Noah Shinn's personal agent, for about $35 a month, replacing a roughly $250,000-a-year stack that includes a …
OpenAI reported that its GPT-6 Astra model drew roughly half as many high-severity misaligned-behavior flags as GPT-5.6 Sol across internal evaluations covering more than 54,000 tasks, according to an…
OpenAI's GPT-6 Astra was the only model to complete Axiom Math's DrivingBench real-world driving test, finishing a 134.7-meter cone course in a 2022 Toyota Corolla in 5 minutes 22 seconds on its secon…
Three AI engineers at the startup Axiom — Aditya Ramabadran, Simon Mahns, and Tobias Gessler — used OpenAI's GPT-6 Astra to autonomously drive a 2024 Toyota Corolla through an In-N-Out Burger drive-th…
OpenAI's GPT-6.1 Sol supports async tool calling, letting an application hand off a slow task or human question while the model continues independent work, but the API only removes the model-side wait…