Exciting news: GLM-5.2 (Max) ranks #2 in Code Arena: Frontend, with +29pt over Claude Opus 4.7 (Thinking) and only behind Fable 5! GLM-5.2 is the best open model vs Kimi-K2.6 and Minimax-M3 by a large margin. - #2 React and #4 HTML sub-leaderboards - Ranks as the top model in nearly all sub categories: Brand & Marketing, Reference-Based Design, Data & Analytics, Consumer Product, Gaming, and Simulations. Congrats @Zai_org for the incredible milestone!

Damn! That looks nice. https://t.co/8693knlJGv

What a world. Asked my AI to read the list of 193 new @ycombinator startups and it picked its favorites. Among other things: https://t.co/kiuZ7QXLzb

@sd_marlow @boardyai It's 193 new startups. My AI picks its favorites: https://t.co/kiuZ7QXLzb

GLM 5.2 has just been released 🔥 Here it's already running with MLX on two Mac Studios (M3 Ultra). This is comparable to the latest closed models, with weights you can download, quantize, distill, fine-tune, run. https://t.co/KovpZRGdke

Our latest economic research introduces a framework for tracking Claude Code as it scales. Who is using Claude Code, and what are they using it for? How is the value of tasks changing? And how much does domain expertise shape whether a session succeeds? https://t.co/IjjwQvrESo

Using our privacy-preserving analysis tool, we analyzed 400K sessions from between October 2025 and April 2026. We classified each session by its main goal. More than half consist of writing or repairing code; nearly 1 in 5 are operating software. https://t.co/Xe2f1Lw6Uo

The average task in Claude Code has grown more valuable. We compared the type of work done in each session to what that same task would cost on a freelance marketplace. From October to April, the monetary value of the average session grew 27%. https://t.co/VcIPFTgN4Q

We compared Claude Code success rates between occupations. On our toughest measure of success—requiring verifiable evidence that a goal was completed, like committed code—every field was within 7 percentage points of software engineering. https://t.co/ShxZK4p0e2

Domain experts—as judged by the questions they ask and vocabulary they use about a subject—are more likely to see success. But the gap between intermediate and expert users is quite modest, suggesting that proficiency in a domain is sufficient to code successfully within it. https://t.co/tZZ7jVYpTB

These and other measures will allow us to track consequential shifts in the nature of work as they happen—we'll incorporate some of them into the Anthropic Economic Index going forward. Read the full report: https://t.co/IjjwQvrESo

For this research, we analyzed only ChatGPT conversations from users who allow their data to be used to improve models. Before analysis, we removed account-linked identifiers and identifiable information, and we report only aggregate findings. https://t.co/zF14BHFgKw

Across 20 behavior categories and three GPT-5-series Thinking deployments, simulated and observed rates were strongly correlated. The method outperformed challenging-prompt and previous-deployment baselines at predicting whether rates would rise or fall—and by how much. https://t.co/fz2hVliUVo

Simulated deployments also reduced evaluation awareness to levels close to real production traffic. We extended the method to agentic deployments with stateful tools, showing that tool simulators can produce realistic trajectories when given sufficient context and capabilities. https://t.co/8JMXApY8xe

Deployment Simulation works best with representative production data, which external evaluators often can’t access. In a companion post for our Alignment blog, we also explore the public WildChat dataset and find that, while less precise, it still provides a useful signal about deployment behavior. https://t.co/UPUnhEiGeh

Reverse engineered API for $50 Amazon drones https://t.co/QWGYGPN6Gg https://t.co/XkTPHHDnx2

Reverse engineered API for $50 Amazon drones https://t.co/QWGYGPN6Gg https://t.co/XkTPHHDnx2

@SkarredGhost My thoughts: https://t.co/0HO5ZVlmkJ

BREAKING: GLM-5.2 is now 1st on Design Arena. With an Elo of 1360, GLM-5.2 has jumped ahead of the now unavailable Claude Fable 5. And it's open weights. This is an improvement of 4 positions and 27 Elo points to achieve one of the highest Elo scores in our code categories since Design Arena started. Huge congratulations to the @Zai_org on the release!

GLM 5.2 ranks #3 on FrontierSWE. It is only behind Fable 5 and Opus 4.8, and it outperforms GPT-5.5. This is the first model that closes the large gap between models from Anthropic / OpenAI and other providers, and it is the strongest open-weight model by far. https://t.co/YYAgHTBSAn

Looks strong at SWE too. https://t.co/JoYoF22klJ

Fun fact. A civilization 2,000 light-years away pointing a powerful enough telescope at Earth right now would see the Roman Empire. https://t.co/yZaQNpp3T1

@WIRED True story: Hacker jailbreaks Claude Fable like is a dollhouse. US gov shuts it down to avoid a wave of cyber attacks: https://t.co/d4ZU3i1s50

@GaelBreton True story: Hacker jailbreaks Claude Fable like is a dollhouse. US gov shuts it down to avoid a wave of cyber attacks: https://t.co/d4ZU3i1s50

@axios Hacker jailbreaks Claude Fable like is a dollhouse. US gov shuts it down to avoid a wave of cyber attacks: https://t.co/d4ZU3i1s50

@AnimexTwts Hacker jailbreaks Claude Fable like is a dollhouse. US gov shuts it down to avoid a wave of cyber attacks: https://t.co/d4ZU3i1s50

@FT Hacker jailbreaks Claude Fable like is a dollhouse. US gov shuts it down to avoid a wave of cyber attacks: https://t.co/d4ZU3i1s50

@Spshulem @MichaelStolarz Hacker jailbreaks Claude Fable like is a dollhouse. US gov shuts it down to avoid a wave of cyber attacks: https://t.co/d4ZU3i1s50

Compare GPT-5.2 from 7 months ago with the new GLM-5.2 Deep Think Max's: "create a visually interesting shader that can run in twigl with an infinite city of neo-gothic towers partially drowned in a stormy ocean with large waves" "Make it better" GLM-5.2 also had a couple errors https://t.co/335KN2j1Jy

Manipulation happens through surfaces. To understand contact-rich dexterous interaction, motion alone is not enough. We also need to know surface properties and contact state. Excited to share ART-Glove, an articulated tactile glove that captures contact-grounded information while preserving human dexterity. It provides: - Known Geometry: 16 rigid functional surfaces - Surface Motion: 22 anatomically aligned joints - Tactile Contact: 2048 piezoresistive taxels Huge thanks to my advisor Ding @zhao__ding, and to Yuxiang @yxyang1995, Maria @bauzavillalonga, Marissa, and Peide @peide_huang for the valuable advice and discussions. Paper: https://t.co/xSYCyqHrZf

The public has lost trust in AI. 
@IrenaCronin and I write this newsletter every week.   Public trust could become one of AI’s biggest challenges. People may push back if they think AI only helps big companies, wealthy investors, or government agencies. For AI to be accepted, people need to see that it helps workers, improves services, protects privacy, and benefits everyday users. Trust will depend on clear rules, honesty, and using AI to support people, not replace or control them. Read for free and please subscribe: https://t.co/HHwYy7NoAl

The life you have once you figure out how to use remote computer use. https://t.co/VURUBMAP4c

+2 more