We've unified plans across Grok, Cursor, and X Premium. Introducing The X Subscription.

Introducing GLM-Legal Built on GLM-5.3-Flash, GLM-Legal matches frontier models on mainstream legal benchmarks. Designed for real legal workflows: contract review, evidence analysis, compliance, and more. Tech Blog: z.ai/blog/glm-legal Apply for API Access: z.ai/consultation/legal Try AutoClaw: autoclaw.z.ai

We're launching our leaderboard for FrontierCode 1.1, a private evaluation from Cognition that tests models on completing coding tasks in a way that would actually get merged by maintainers of real-world repositories @cognition announced FrontierCode in June 2026 and updated it to version 1.1 in July. Cognition built the dataset with open source maintainers who spent more than 40 hours per task. Each task requires the agent to make a change to a real open source repository, then checks both unit tests and qualitative tests for what maintainers would require to merge the change. We run FrontierCode 1.1 Main, the primary 100-task set of the benchmark's 150 tasks, using our Stirrup open source reference harness for apples-to-apples comparisons across models. All tasks are private to avoid contamination. Each attempt is graded with both blocking and non-blocking criteria. An attempt that fails any blocking check scores zero, and one that passes them all scores between 70% and 100%. Unlike other coding evaluations, FrontierCode has a more holistic view of success based on real usage. Models need to pass unit tests, but are also assessed on behaviors like making excessive changes and large diffs, modifying files outside the relevant scope, or adding unhelpful tests. We think this approach is valuable in the current context: models are now very capable at coding tasks, but their changes can still require substantial cleanup before a maintainer would merge them. Key takeaways: ➤ Leading models still fail more than 40% of the time: Claude Opus 5.5 (max with fallback) leads at a score of 52% after passing blockers in 58% of its attempts. Nine of the 100 tasks were not passed in any attempt by any of the models we evaluated for launch ➤ The top of the leaderboard is tightly packed: nine models from six labs score between 43% and 52%. MiMo-V2.6-Pro is the leading open weights model at 45%, followed by DeepSeek V4.1 Flash (max) and Kimi K3 (max) at 44% ➤ GPT-6.1 Sol (max) offers the strongest performance for its cost: At $0.81 per task it scores 49%, matching Claude Fable 5.1 (max with fallback). This means it achieves similar performance to leading Anthropic models, at less than a tenth of the cost per task Thank you to @silasalberti, @ybenpan, @cl571128, and the rest of the team at Cognition for building FrontierCode and collaborating with us to bring it to Artificial Analysis - we’re working to add FrontierCode to our coding agent leaderboard very soon.

cloudflare is giving startups up to $350k in credits 💸 most people will see the big number and scroll. don't. the tier most of you can actually get is $10k the $10k tier: - bootstrapped or self-funded - under $1M raised what it covers (valid 1 year): - hosting for your app, backend and databases - file storage - ai on open models (capped at $2.5k on this tier) - websites, images and video free on every tier: 3 enterprise domains, ddos + bot protection, cdn, ssl building ai apps or agents? that covers a lot of your hosting for a year bigger tiers ($100k / $350k) need a partner vc behind you who qualifies: - incorporated in the last 10 years - for-profit, building a tech product - live website + business email on your own domain (no gmail) - active linkedin, x or github - first time applying review takes up to 48 hours the fine print: 1. you need a card on file. when the year ends or credits run out, they bill it 2. no extensions 3. no claude here. open models only, and ai gateway isn't covered right now 4. domain registration isn't covered 5. go over a product cap and you pay the overage 6. agencies, consultancies and personal blogs aren't eligible claude startups gets you the model. cloudflare gets you the infra to host it what would you build first with $10k of cloudflare credits?

Jack Doherty might have made the biggest mistake of his life Jack (@dohertyjackk) signed a sponsorship deal with Shuffle (a crypto casino) but didn’t fully read the contract One of the terms is a 6 month ban on posting online He said he thought it was a joke, but Shuffle is enforcing it Shuffle CEO (Noah Dummett) just confirmed it’s real and said they’ll stick to the contract even though Jack lawyers are pushing back Wild times

Grok Bot now has its own email. Bot can use it to sign up for services, contact businesses for you, or schedule time with someone.