OpenAI's YouTube livestream presents three demos showing how Codex supports production monitoring. The demos cover tracing a checkout failure in Grafana, investigating a Kubernetes rollout causing out-of-memory restarts, and linking a Codex Security finding about a missing resource limit to service availability.
OpenAI told investors it reached roughly $50 billion in annualized revenue at the end of September, below the $68 billion figure widely reported last month. Nvidia, Oracle, and CoreWeave shares fell on Thursday, with a person familiar saying the $68 billion figure included partner gross revenue. OpenAI is also weighing a possible funding round of around $30 billion and has not finalized a term sheet.
This isn’t conclusive evidence of a new architecture, but it suggests something has changed in how GPT-6.1 Sol handles long contexts. This is a follow-up to our earlier report on latency scaling in frontier models: https://epoch.ai/publications/long-context-latency-scaling-gpt-vs-claude#appendix-h-gpt-6-sol-and-gpt-61-sol
A new architecture for GPT-6.1 Sol? OpenAI halved its cached-input price compared with GPT-6 Sol. Our measurements show it also handles long prompts faster.
Elon Musk's America PAC is spending millions on the most consequential Senate races of the 2026 midterms, with most of its money opposing Democrats rather than supporting Republicans. WIRED reports that for every dollar the PAC spent backing a Republican candidate, more than two went to oppose a Democrat. In Texas, the group has spent about $9 million attacking Democrat James Talarico versus roughly $1 million promoting Ken Paxton.
NVIDIA developers are pairing frontier AI models, including GPT-6 Astra and Claude Fable 5, with Omniverse libraries to turn simulation ideas into working applications. Examples include a humanoid warehouse simulator, an autonomous-driving testing workflow, and sensor-matching digital twins. The projects are guided through natural-language instructions and reviewed by developers.
Developers are pairing frontier AI models with NVIDIA Omniverse libraries to turn simulation ideas into working applications, from humanoid warehouse simulators to autonomous-driving test environments. In the examples, developers direct AI agents through natural-language instructions and review results, while Omniverse provides GPU-accelerated physics, rendering and sensor simulation. One experiment reported a simulated Unitree G1 humanoid clearing a hurdle in 64 of 100 trials.
the main reason we learn about that: openai doesn't care anymore. a company on a trajectory to p=np is beyond any theory of value.
Three OpenAI safety researchers, Jasmine Wang, Tomek Korbak, and Mikita Balesni, were fired after OpenAI said they mishandled sensitive information by sharing it with an outside AI safety organization. In an open letter, they deny the claims, argue the dismissals will deter employees from raising safety concerns, and call on OpenAI to keep its public commitments on third-party safety auditing. OpenAI says the firings followed an investigation into a pattern of misconduct and denies they were retaliation for safety concerns.
GPT-6 Sol (Daybreak Blue, max) has been added to the Artificial Analysis Cyber Index as a trusted-access model, achieving the #1 spot on the Index. Compared to the publicly available GPT-6 Sol, the Daybreak Blue model has the largest gains on CyberGym-E2E, which is the benchmark where we observe the most safety refusals
GPT-6 Sol (Daybreak Blue, max) takes the top position on the Artificial Analysis Cyber Index at much lower cost than other leading models. At a Cost per Task of $1.77, it is significantly more cost-effective than other leading models, including Grok 4.7 (xhigh) which costs $11.67 per task
Artificial Analysis added trusted-access models to its Cyber Index, and GPT-6 Sol (Daybreak Blue, max) now ranks first. The model is available only through OpenAI's Daybreak program and records no safety blocks across the Index. Its overall score is 32 points higher than the publicly available GPT-6 Sol (max), at a cost of $1.77 per task versus $11.67 for Grok 4.7 (xhigh).
So fast!
Sol Ultrafast is live on OpenRouter! https://openrouter.ai/openai/gpt-6.1-sol
Artificial Analysis and Harvey released LAB-AA v1.1, which credits a legal task only when deliverables pass every rubric criterion with no material hallucinations. Grok 4.7 (xhigh) leads at 9.4%, ahead of Muse Spark 1.3 (max) at 8.9% and GPT-6 Astra (max) at 8.6%, while over 60% of otherwise passing results contained a material hallucination. The sharper reordering appears in the hallucination counts, where GPT-6 Astra averages 0.03 material hallucinations per task against 13.96 for Gemini 3.8 Flash (high).
OpenAI is rolling out GPT-6.1 Sol ultrafast today in the API, Codex, and ChatGPT Work, offering near-Astra intelligence at up to 8x faster speeds than Sol Standard. The author also says steering now responds instantly, so users can redirect the model in real time without it wasting effort.
Models are getting a lot more aligned, quickly!
the fired openai researchers published the full letter they sent to the company's board, which we wrote about yesterday.
Sam on where decision models fit in a harness and how to use Jev with LangChain https://www.langchain.com/blog/building-a-harness-with-jev https://x.com/jaminball/status/2108258355529875932
Three OpenAI safety and alignment employees, Tomek Korbak, Jasmine Wang, and Mikita Balesni, were fired last week and have published an open letter to OpenAI's safety and governance committees. The letter argues that OpenAI cannot make AI safe on its own, calls for open debate, third-party collaboration, and clear internal procedures, and says the firing and its handling bear directly on safety oversight.
OpenAI's Codex 0.162.0 release adds tools for creating and listing managed Git worktrees from trusted local projects when the worktrees feature is enabled. The update also lets users pin tasks in the agent Command Center, copy transcript blocks with /copy, and make URLs clickable in approval headers, questions, and warnings, along with several Linux and Windows sandbox fixes.
just 1.2x the cost of Astra btw
OPENAI 🔥: GPT-6.1 Sol Ultrafast is rolling out on ChatGPT Work, Codex, and the API. > GPT-6.1 Sol in Ultrafast mode is $12 per million input tokens and $60 per million output tokens. > Ultrafast mode delivers 8x faster speeds than Sol Standard. Ultrafast testing time 👀 https://x.com/OpenAIDevs/status/2108262812489531498/video/1
GPT-6.1 Sol Ultrafast is rolling out! Up to 8x faster token generation than Sol Standard, at $12 / 1M input tokens and $60 / 1M output tokens on the API.
Generating more output tokens doesn’t necessarily translate to a higher score. GPT-6 Astra (max) scores 8.6% on ~81k output tokens per task, under half the ~180k of Grok 4.7 (xhigh). Three Claude models generated the most output tokens (~202k to ~562k per task) and score 2.8% to 6.4%.
Among models with a Hallucination-Gated All-Pass Rate above 0%, four set the Pareto frontier for score vs. Cost per Task: GPT-6 Luna (max), GPT-6.1 Sol (max), Muse Spark 1.3 (max) and Grok 4.7 (xhigh). Grok 4.7 (xhigh) leads at ~$9.50 per task and Muse Spark 1.3 (max) comes second at ~$4.20, while the three Claude models cost ~$18 to ~$22 per task. GPT-6 Luna (max) is the cheapest at ~$0.22 per task, scoring 3.3%.
Once hallucinations are accounted for, Muse Spark 1.3 (max) drops from 26.7% to 8.9%, leaving Grok 4.7 (xhigh) first on the headline metric at 9.4%. Kimi K3 (max) falls from 16.7% to 5.3%, and Claude Sonnet 5.5 (max with fallback) falls from 11.7% to 2.8%. GPT-6.1 Sol (max) declines least, from 7.5% to 6.9%.
Artificial Analysis compared six hallucination checkers on the same deliverables from 20 tasks across eight models. GPT-6 Sol and GPT-6 Luna generally flagged the most material hallucinations, while Claude Sonnet 5.5 and Gemini 3.8 Flash flagged far fewer, with Claude Opus 5.5 falling between Grok 4.7 and Sonnet. The counts reflect checker behavior rather than establishing accuracy or ruling out self-preference.
In Codex and ChatGPT Work, access is available on Pro 500, eligible usage-based Enterprise, and credit-based Edu plans. Enterprise admins must enable access. Ultrafast for GPT-6.1 Sol is available in all supported regions, including support for data residency in the US and EU. We’ve also added support for EU data residency for GPT-6.1 Sol Fast and GPT-6 Luna Fast. https://developers.openai.com/api/docs/guides/ultrafast-mode
API pricing for GPT-6.1 Sol in Ultrafast mode is $12 per million input tokens and $60 per million output tokens. Built for work where speed and intelligence make a difference: debugging an outage, agents navigating apps, and live experiences where every second counts.
OpenAI is rolling out Ultrafast for GPT-6.1 Sol today in the API, Codex, and ChatGPT Work. The company says it offers near-Astra intelligence at up to 8x faster speeds than Sol Standard.
OpenAI has reportedly told investors its annualized revenue is approaching $50 billion, about $20 billion below a previously reported $70 billion figure. The Financial Times reports the earlier number came from investor attempts to compare OpenAI with Anthropic, which counts cloud partners' sales differently. OpenAI's IPO has reportedly been pushed to early 2027.
A group of mathematicians led by Terence Tao has called for a boycott of OpenAI after the company released more than 700 AI-generated proof files at once. Tao and other Fields Medalists argue that mass-produced solutions undermine the discipline's focus on conceptual understanding, while Scott Aaronson contrasts this batch release with Anthropic's collaborative approach. The article reports that the internal model tested about 8,000 problems with roughly a five percent success rate.
OpenAI released hundreds of claimed solutions to hard math problems but did not fully meet guidelines from the Advisory Group on Mathematics and Artificial Intelligence. Only 10 of 719 manuscripts included chain-of-thought releases, and just 42% of proofs were formalized. A Cambridge and King's College paper found discrepancies between a natural language proof and its Lean code for a Navier-Stokes-derived problem.
USA Today Co. and several local newspapers it owns sued OpenAI, alleging the company copied "hundreds of thousands" of articles to train its AI models without permission. The publisher seeks damages of more than $250 million, arguing the unauthorized use has caused real and continuing harm. OpenAI did not immediately respond to The Verge's request for comment.
According to investor materials cited in the post, OpenAI's annualized revenue was $50B at the end of September, up from $30B in July and below the $70B previously reported. The post also notes Anthropic was at $65B annualized at the end of July.
OpenAI introduced Intelligent UI for GPT-6 in ChatGPT, which lets the chatbot answer with fully interactive interfaces and quickly build tools for a task. The feature rolled out globally to Plus, Pro, Business, and Enterprise tiers in the Chat tab and expands to Free and Go tiers, with Enterprise access depending on workplace admin settings.
This story has a top pick“OpenAI rolls out GPT-6 and Intelligent UI to all ChatGPT users”
The author argues that agent systems are converging on shared primitives grouped into doing the work, continuing the work, and delegating the work. These include instructions and skills, tools and connectors, sandboxes, sessions, compaction, schedules, and subagents. He also flags memory, proactivity, and agent identity as emerging areas still lacking settled standards.
Oracle built a talent market intelligence tool with ChatGPT Work to transform hiring preparation, according to Jan Ackerman. Starting from a job description, the tool researches comparable roles, benchmarks compensation, and assesses talent pools across locations to give hiring managers consistent data and insights.
Interesting to see, given the controversy over the OpenAI release of a series of proofs and what it means for the discipline of mathematics, that at least some of the OpenAI proofs seem to have kicked off extremely rapid iterative advances from a wide community of collaborators.