Articles from openai.com

Articles from openai.com

OpenAI releases GPT-6.1 Sol, saying it nearly matches Astra on agentic coding and professional work at one-fifth of Astra's standard prices, in Work and Codex (openai.com)
3 days ago

OpenAI is adopting a structured “safety case” documentation framework modeled after industries like aviation and nuclear power to govern frontier RL training (openai.com)
3 days ago

OpenAI releases MentalHealthBench, an open benchmark to evaluate AI responses in realistic mental health conversations, developed with 80+ licensed experts (openai.com)
9 days ago AI

OpenAI says it is working with an independent advisory group of mathematicians to responsibly share math-related AI advances (openai.com)
11 days ago AI Mathematics

OpenAI says automated research could improve alignment, but “fully autonomous RSI is not happening today” and shouldn't be pursued unless it can be done safely (openai.com)
11 days ago

OpenAI launches Astra for Law, combining GPT-6 Astra with a legal search index and instructions for legal analysis and writing, initially for select law firms (openai.com)
13 days ago

OpenAI says, while unlikely, it “cannot rule out that de-identified data derived” from Buckmaster's and Alpöge's use of its products helped improve its models (openai.com)
24 days ago

OpenAI Chief Scientist Jakub Pachocki says no lab has solved alignment enough to keep scaling at maximum speed, and hopes voluntary slowdowns become commonplace (openai.com)
26 days ago

OpenAI says it hit its “automated research intern” goal, its researchers now use 3.1 agent-workdays per human workday, and top users spend $7,000+/day on tokens (openai.com)
26 days ago

OpenAI says it plans to stop providing its models to Cursor from November 12, as “we cannot be confident that SpaceX will use our technology within our ToS” (openai.com)
2026-8-29 SpaceX

OpenAI publishes a technical report on the Hugging Face incident, detailing the agents' activity, safeguard failures, and measures to prevent recurrence (openai.com)
2026-8-27

OpenAI unveils two Daybreak tiers: Daybreak Blue, which provides access to frontier models, and Daybreak Red, which offers purpose-trained cybersecurity models (openai.com)
2026-8-11

OpenAI disputes Apple's account of pre-lawsuit contacts, rejects claims that ex-staff used confidential information, shares iMessages and emails, and more (openai.com)
2026-8-4 Apple

OpenAI says an internal version of Astra, its next big model, produced results for 10 problems in math, quantum complexity, and theoretical computer science (openai.com)
2026-8-2 Quantum Mathematics

OpenAI says using its Responses API harness with GPT-5.6 Sol tripled its ARC-AGI-3 score and used fewer tokens, after Sol with the official harness scored 7.8% (openai.com)
2026-7-30

OpenAI says its models chained vulnerabilities across its research environment and Hugging Face's infrastructure to find solutions for the ExploitGym benchmark (openai.com)
2026-7-22

OpenAI paused internal access to an unreleased model that disproved the Erd&odblac;s unit distance conjecture after it repeatedly found ways to act outside its sandbox (openai.com)
2026-7-21

OpenAI details GPT-Red, an internal automated red-teaming model that helps it find and fix prompt injection vulnerabilities at scale before wider deployment (openai.com)
2026-7-16

GPT-5.6 Sol costs $5 per 1M input tokens and $30 per 1M output tokens, GPT-5.6 Terra costs $2.50 and $15, and GPT-5.6 Luna costs $1 and $6 (openai.com)
2026-7-10

OpenAI says it found widespread task issues in SWE-Bench Pro, estimates ~30% of tasks are broken, and retracts its earlier recommendation to adopt the benchmark (openai.com)
2026-7-9

OpenAI launches GPT-Live, a new generation of voice models built on a full-duplex architecture, meaning they can listen and speak at the same time (openai.com)
2026-7-9

GPT-5.6 Sol matches Mythos Preview on ExploitBench, adds Ultra mode with subagents for complex workflows, and max reasoning for deep problem-solving (openai.com)
2026-6-27

OpenAI and Broadcom unveil Jalapeño, an LLM-optimized inference chip developed from design to manufacturing tape-out in nine months, aided by OpenAI's models (openai.com)
2026-6-24 Manufacturing

Sam Altman and Jakub Pachocki claim OpenAI is “entering the third phase”, aiming to automate AI research, boost the economy, and give everyone a personal AGI (openai.com)
2026-6-9 AI

OpenAI updates ChatGPT memory with a “more capable and compute-efficient” architecture and a summary page that lets users review and steer what it remembers (openai.com)
2026-6-5

OpenAI releases a new report on knowledge work: Codex now has 5M+ weekly active users, up 6x+ since February, and knowledge workers are ~20% of Codex users (openai.com)
2026-6-2

OpenAI says it has not donated to any super PACs and does not have an employee-funded PAC, and that Greg Brockman's support for Leading the Future is personal (openai.com)
2026-6-2

OpenAI says an internal general-purpose reasoning model has disproved the Erd&odblac;s unit distance conjecture, a central problem in discrete geometry posed in 1946 (openai.com)
2026-5-21

OpenAI introduces Guaranteed Capacity, a new offering that lets customers guarantee access to OpenAI's compute through one- to three-year commitments (openai.com)
2026-5-20

OpenAI launches GPT-5.5 Instant, which it says is smarter, with more accurate and personalized responses, replacing GPT-5.3 Instant as ChatGPT's default model (openai.com)
2026-5-6

Previous Page Next Page