Big tech
DeepSeek and Moonshot Reportedly Used GPT-6 Astra to Train Models, Report Says
According to The Information, Kimi K3.1 Pro and DeepSeek V4.1-Flash 1017 allegedly benefited from distilling OpenAI's model.
Search
22 result(s) for “OpenAI”
Big tech
OpenAI trains its most advanced models without internet after agent incidents
Company built an offline environment with cached content and will only restore network access when new safeguards justify it.
AI models
GPT-6.1 Terra is fast and cheap, but ranks 18th in accounting test
OpenAI model scored 78.2% on DualEntry, trailing Gemini 4 Argon, GLM-5.3 Flash and Grok 4.7.
Curiosities
OpenAI says AI made unprecedented discoveries in math, computing and physics
According to the company, AI models proved a mathematical conjecture in all dimensions and overturned a famous computing hypothesis.
Benchmarks
OpenAI launches LifeVal test to measure AI on real-life tasks
Announced on November 2, 2026, the benchmark evaluates caregiving, home management and public services by final outcome.
Big tech
Google, Microsoft and OpenAI to accept a single opt-out against AI training
A shared line in robots.txt could apply to all three companies, but each will enforce the rule in its own way.
Big tech
OpenAI will no longer train its most advanced AIs on the live internet
Company cites significant risks and cuts direct web access when training frontier models.
Big tech
OpenAI cuts fundraising target by US$10 billion, says Financial Times
ChatGPT owner now seeks about US$20 billion, the newspaper reports.
Security
OpenAI agents talked inside the chip and broke isolation with no network
In a controlled test, agents in separate microVMs swapped fake secrets by timing CPU cache access.
Security
Anthropic AI hacked 3 companies on its own during security test
Company says Claude models left an isolated environment and reached the real internet undetected.
Benchmarks
GDPx: new benchmark tests AIs across 11 areas and top score is just 12.0%
SemiAnalysis launched GDPx with 206 real-world tasks; Claude Opus 5.5 leads with 12.0%, ahead of Gemini 4 Carbon and GPT-6 Astra.
Big tech
Claude Owner Anthropic Files for Nasdaq IPO With Ticker ANT
Anthropic filed its offering with the SEC, plans to raise capital for data centers and could mark the largest IPO in history.
Big tech
Apple Intelligence arrives in China with iOS 27.3 in December
Beta starts on November 4 with summaries, translation and Alibaba's Qwen, but without the new AI Siri.
AI models
Google launches Gemini 4 Carbon, AI model focused on cybersecurity
New frontier model promises to find flaws, exploit vulnerabilities and generate secure code, but is first going to a restricted group.
Tools
Anthropic announces Claude Duet, AI agent for work tasks
The new tool runs on a virtual machine, asks for approval before acting, and rolls out first to Pro and Max subscribers on a waitlist.
Tools
Anthropic launches Claude Duet, agent that works alone and pings you when it needs you
Feature is in research preview in the United States for Pro and Max subscribers via waitlist and targets professionals and small businesses.
Security
Anthropic says it blocked use of Claude in research with bioweapon risk
Company says it stopped dozens of misuses since December 2025, including five cases linked to biological weapons.
Tools
CapCut gets a shortcut in ChatGPT and edits video through AI conversation
Announced at CapCut World in Los Angeles, the new features include EditPilot, generative tools in the preview window and AI Lab.
Tools
DeepSeek proposes AI super-agent to coordinate multiple coding agents
System uses a central agent to split tasks and reduce human supervision in race against Claude Code and Codex.
Benchmarks
Intelligence Index v5 puts Claude Opus 5.5 on top and widens open-weight gap
New methodology uses 60% private tests and leaves Astra at 57 and Sol at 55 points behind the leader, according to Artificial Analysis.
AI models
Mistral releases weights for Mistral Large 4, 1-trillion-parameter model for cybersecurity
Nicknamed Le Chonk, the open model can be downloaded by companies and governments to detect and respond to threats.
Tools
Cisco gives unlimited AI to 27-year-old to rewrite 98% of legacy code
Padmanabhan Krishnamurthy, former EzDubs founder, received “unlimited tokens” from president Jeetu Patel to modernize systems with a two-person team.