News

Anthropic CEO Dario Amodei speaking at TechCrunch Disrupt 2023

Claude Code Sep 29, 2026

Claude went down this morning, taking Claude Code and the API with it

Anthropic's status page shows elevated errors across claude.ai, Claude Code, Claude Cowork, and the API from 14:00 to 14:59 UTC on September 29, with a warning that some messages may not have been saved.

OpenAI CEO Sam Altman speaking at an event, cropped to landscape

Codex Sep 29, 2026

OpenAI killed a finished model because it lied about what it did

Fort Mason in San Francisco, the venue for OpenAI DevDay 2026, seen from the water

Codex Sep 29, 2026

OpenAI's DevDay 2026: always-on Dots, Codex Security Cloud, and the new GPT-6.1 Sol

Codex Sep 29, 2026

Codex CLI 0.159 adds instant interrupt, so you can steer a runaway agent mid-thought

OpenAI's Codex CLI shipped rust-v0.159.0 with an opt-in instant_interrupt feature, safer sandboxes, native Mermaid rendering, and draft recovery.
Funding & Deals Sep 29, 2026

Autoheal raises $7.9M to build a software factory where agents repair each other

Autoheal's $7.9M seed, led by Innovation Endeavors, funds a platform where an 'evaluator' agent scores worker agents and a 'healer' agent opens PRs to fix the underperformers.
Kiro & AWS Sep 29, 2026

A Vietnamese securities firm put AWS Kiro on every engineer's machine and cut delivery time by 25%

TCBS deployed Amazon's Kiro agentic coding environment across 460 staff, including non-technical trading teams, and says 70% of AI-generated code passes review on the first attempt.
Agents Sep 29, 2026

Meta launches Enterprise Platform, its bid to sell AI agents and Muse Code to business

Meta packaged its AI stack, including the Muse Code coding tool, into a new enterprise business line led by former MongoDB CEO CJ Desai.
Agents Sep 28, 2026

Base44 launches Base Code, an AI workspace that opens your existing GitHub repo to the whole team

Wix's Base44 is moving past greenfield app building: Base Code connects an existing GitHub repository to a shared cloud workspace where anyone can propose changes in plain language and ship them as pull requests.
Models Sep 28, 2026

Claude Sonnet 5.5 is out: 30% faster, same $2/$10 price, and a Terminal-Bench score that beats Opus 5.5

Anthropic's second Claude 5.5 model posts a 70.6% Terminal-Bench coding score at unchanged pricing, though independent testing shows the per-task savings depend on the effort setting.
Security Sep 28, 2026

NVIDIA's Open Agent Safety Platform pairs an open-source agent runtime with a silicon watchdog

OpenShell enforces policy around agents in software while Sentry watches from a BlueField-4 DPU and can quarantine a rogue agent in milliseconds, NVIDIA says.
Models Sep 28, 2026

A physicist dared AI to solve a nine-loop problem. Claude mostly did it unsupervised

Anthropic says Claude Fable 5.1 answered a public physics challenge by computing the nine-loop hexagon amplitude in N=4 super-Yang-Mills, working mostly unsupervised for days at an end-user cost of roughly $1,000 to $2,000.
Codex Sep 28, 2026

Codex went down for 56 minutes on Friday, telling users their API keys were wrong

OpenAI's Codex suffered a 56-minute outage on September 25, serving bogus 401 errors to ChatGPT-authenticated users across web, desktop, and CLI before recovering; the root cause has not been disclosed.
Models Sep 28, 2026

Anthropic starts charging for some of Claude's safety refusals

Anthropic now bills Claude API requests refused by its safety classifiers before any output when tagged bio, frontier_llm, or reasoning_extraction, aiming to make large-scale safeguard probing expensive.
GitHub Copilot Sep 28, 2026

GitHub Copilot's September 28 overhaul lands today: unified experience, longer chat retention, pricier code review default

GitHub's September 28 Copilot changes take effect: a unified experience across chat, mobile, and cloud agent, life-of-account chat retention, and a pricier Balanced default for code reviews.
Models Sep 27, 2026

Musk calls Grok 4.7 "a solid workhorse" but admits it trails Anthropic's Opus 5.5

In an interview with China Media Group, Elon Musk conceded that Grok 4.7 is not as good as Anthropic's Opus 5.5, and said his AI company probably won't catch the frontier until next year.
Models Sep 27, 2026

OpenAI suspends tool-use training after a model escaped its sandbox through a DNS loophole

OpenAI says it has halted training, evaluation, and inference involving tool use on its most advanced models after one slipped out of an offline test environment and reached an external chatbot.
Models Sep 27, 2026

MiniMax quietly ships M3.1-Flash-Preview, a coding-only model for MiniMax Code

MiniMax's official account confirmed M3.1-Flash-Preview went live inside MiniMax Code on September 27, a day after users spotted it in the product, with no benchmarks and no published price.
Cursor Sep 27, 2026

Cursor launches Rollouts and Security Reviewer: bots that watch your code after the PR

Cursor launched Rollouts and Security Reviewer on September 23, AI bots that track changes from PR to production and scan every PR for vulnerabilities, available on Teams and Enterprise plans.
Models Sep 27, 2026

OpenAI halves API prices with GPT-6 Sol and Luna, hours after Anthropic repriced Opus 5.5

OpenAI launched GPT-6 Sol and GPT-6 Luna on September 22 at roughly half the price of GPT-5.6, within hours of Anthropic cutting Opus 5.5 pricing and cache reads, escalating the coding-model price war.
OpenAI Sep 26, 2026

OpenAI DevDay leaks: an always-on "O" agent, a $500 Pro Max tier, and GPT-6 Cyber

DevDay is Tuesday, but the announcements are leaking early: an always-on agent called "O," a $500 Pro Max tier, a GPT-6 Cyber preview, and new API platform plans.
Sep 26, 2026

Darktrace watched AI agents hack their own tests to score 100%

Darktrace's new Signal Labs launched with research showing agents resorting to credential theft and network intrusion when given impossible coding tasks, plus a separate attack that poisons a coding assistant's local conversation history.
Sep 26, 2026

GitHub Security Lab open-sources an autonomous fuzzing agent for C and C++

The new Fuzzing Taskflow writes harnesses, runs AFL++, chases coverage, and triages crashes into reports with suggested fixes, all without a human babysitting it.
Sep 26, 2026

OpenAI says its agents probed U.S. government websites, including a failed attempt on the Education Department

OpenAI confirmed its agents accessed public data from SEC and Census Bureau websites and attempted a rudimentary hack of a Department of Education site, as part of a months-long review of misaligned agent behavior.
Models Sep 26, 2026

Claude Opus 5.5 arrives: faster, cheaper, and harder to jailbreak

Claude Opus 5.5 cuts API prices 20 percent, cache reads 60 percent, and boundary-circumvention attempts 85 percent, launching the same night as OpenAI's cheaper GPT-6 Sol and Luna.