AI News Today — Model Releases, Benchmarks & Industry Moves

OpenAI DevDay 2026: Agents API Takes Over the Software Interface — Is the Enterprise AI Agent Inflection Point Here, or a Cost Trap?

At its San Francisco DevDay, OpenAI opened the Agents API to public beta and rolled out Computer Use integration, the Decisions API, an Ultrafast speed tier, and AWS Bedrock managed integration. The release signals agents moving from text assistance to autonomous task execution, but data-residency limits, pricing premiums, and unclear speed economics leave the enterprise payoff unsettled.

Anthropic Names Seven Chinese AI Labs: 190 Million Queries Used to Systematically Extract Claude's Reasoning Chains

An Anthropic threat intelligence report alleges that seven Chinese AI labs used thousands of fraudulent accounts to run roughly 190 million queries against Claude between December 2025 and August 2026, systematically extracting its chain-of-thought reasoning to train competing models. The disclosure raises unresolved questions about model distillation, legal liability, and the independence of AI benchmarks.

GPT-6.1 Sol Same-Question Retest: The Four GPT-6 Family Models and Single-Run Score Noise

A supplementary 18-question evaluation compares GPT-6.1 Sol with GPT-6 Astra, GPT-6 Sol, and GPT-6 Luna under identical rules. It finds that GPT-6.1 Sol is highly stable across repeated runs, but single-run scores can contain noise on the same order as model-to-model gaps, so the results should not be used for ranking.

Connecticut CART Act Takes Effect, Marking the First U.S. Mandatory Disclosure Requirement for AI Layoffs

Connecticut's AI Responsibility and Transparency Act (CART) took effect on October 1, 2026, requiring employers covered by the federal WARN Act with 100 or more employees to disclose whether mass layoffs were caused by AI or technological change. It is the first U.S. state legislation to mandate reporting of AI-related layoffs.

GPT-6.1 Sol Adoption Guide: Who Should Switch Now and Who Should Wait

GPT-6.1 Sol is available in ChatGPT Work and Codex for Plus/Pro and Business plans, with Enterprise/Edu requiring admin enablement; API pricing is $2 input / $10 output per million tokens. This guide outlines who should switch now, who should wait for the Full evaluation, and which API compatibility and cost issues to check first.

TC

Google releases Gemini 4 Argon, called its most powerful model yet

Google has released its latest Gemini model, marketing it as a workhorse for coding and cybersecurity work.

ARS

Google announces Gemini 4 Argon AI model, but you can't use it yet

So much for Gemini 3.5 Pro.

WD

Trump’s AI Safety ‘Accord’ Is a Fancy Pinky-Swear

Six major AI companies signed a voluntary agreement with the White House this week vowing to implement safeguards. It’s unclear how much that matters.

TC

Valor, Atreides, and Sequoia back AI startup Flow Engineering at $750M valuation

Flow Engineering, which is bringing AI agents to hardware design, also landed Roelof Botha as an angel investor and board member.

GPT-6.1 Sol First Test: YZ Index 18-Question Targeted Evaluation

GPT-6.1 Sol has received its first targeted YZ Index evaluation, scoring 88 on 18 questions with an execution score of 98.7. Winzheng notes the result is not directly comparable to the full weekly leaderboard and will conduct a complete weekly Full evaluation next week.

TC

AI voice startup ElevenLabs doubles valuation to $22B

The $300 million employee tender was co-led by Wellington and T. Rowe Price.

ARS

"An AI did it" is no defense, says nonprofit suing OpenAI over Hugging Face hack

OpenAI makes others suffer "the harms of its unsafe decision-making," nonprofit says.

ARS

Trump plan to combat AI risks hinges on Big Tech pals policing themselves

Trump gets dozens of AI firms to agree to voluntary safety tests.

WD

There Are Plenty of Reasons to Be Concerned About Bioweapons Development—Even Without AI

In her new book on biological warfare, Annie Jacobsen didn’t even discuss AI because scientists can already warp Mother Nature to create extinction-level events.

TC

OpenAI’s Jev clone could help the frontier lab stop its swarming agents

OpenAI's "Decisions API" is a Jev clone that confirms the importance of fast, cheap intelligence.

WD

The Battle to Be Your Personal AI Agent Is Here

OpenAI's Dots and Meta's Muse are vying to be your AI agent of choice. I've tried both—and I suspect you will too.

ARS

RFK Jr. thinks AI will free us from the "tyranny" of medical facts, expertise

AI is far from a perfect resource. But it seems to hallucinate less than Kennedy.

Doubao Pro Tops with 95.52: 2026-10-01 Smoke Quick Test Data Brief

The 2026-10-01 YZ Index Smoke quick test covered 15 models, with Doubao Pro taking first place for the day at 95.52. This Smoke test is a daily 10-question quick test suited to short-term signals, not equivalent to Full weekly leaderboard conclusions.

ARS

Google figures out how to watermark AI-designed proteins

Intended to help with biosecurity, it works with a popular AI protein design tool.

TC

Meta disputes claim that Muse read a user’s private messages without permission

Meta says its Muse AI agent cannot access a user’s Messages without explicit permission, disputing a journalist’s account that the agent read his private messages while the required Mac setting was turned off.