Introducing: Browser Use CLI 3.0 🌐 Turn any model into a SOTA browser agent > Direct CDP control via browser-harness > Run on cloud browsers or real Chrome > 6× smaller, fewer tokens Try it now 🔗↓

GLM 5.2 ⚔️ Opus 4.7 ⚔️ GPT 5.5 ⚔️ Minimax M3 Should you use open-weight models? We're making a QA benchmark. Using Browser Use v4, we benchmarked these models > Send a task from the LLM Arena dataset to each > Generate the website, score with human evaluation Which model did the best? Check it out ↓

Opus 4.7 ⚔️ GLM 5.2 We're benchmarking models on frontend design. We run each model on Browser Use v4 > One prompt from the public LLM Arena dataset > Generate websites and score them Who won? Check it out ↓

Use the QA skill in Browser Use v4. Your agent builds the app. Give Browser Use the URL and let it: > Test the flow and catch bugs > Click around like a user and judge the UI > Tell your agent what to fix Close the loop. Try it now ↓

We're #1 on Browser Arena! 🥇

GLM 5.2 just beat Fable 5 at website design. 🌐 The crazy part: GLM is text-only. It can build the site, but it can’t inspect the result. So we paired it with Browser Use v2 multimodal QA subagents. > Reviews the site and finds bugs > Judges the aesthetics > Sends GLM targeted fixes to iterate on Build full QA cost less than $0.75. Try it now ↓

Open-weights models have officially caught up We tried GLM 5.2 in BrowserCode > Near Opus-level score > Cheapest model to date This task took $0.18 Try it now ↓

Make full slideshows in seconds using BrowserCode! 🧑‍🎓 Our v4 agent creates photos, slideshows, and videos! > Analyzes the content to gain context > Creates the media file wrt the content > Users can download from the agent's session All it took was one prompt! Try it now ↓

This is how easy it is to deploy B, your browser agent!