Alibaba Qwen3.8-Max-0902 Tops Code Arena WebDev and Cost-Performance Frontier
Alibaba's Qwen3.8-Max-0902 has reached #1 on Code Arena's WebDev leaderboard with 1,691 points and leads the Pareto frontier at a blended $5/MToken, surpassing…
Alibaba's Qwen3.8-Max-0902 has reached #1 on Code Arena's WebDev leaderboard with 1,691 points and leads the Pareto frontier at a blended $5/MToken, surpassing…
Alibaba's Qwen3.8-Max-0902 becomes the top-ranked model on the CodeArena WebDev leaderboard with 1691 points, surpassing Claude Opus 5 and Kimi K3, while mainta…
Alibaba's Qwen model has been integrated into CommandCodeAI, expanding the coding assistant's AI model options.
Qwen3.8-Max has been upgraded to Qwen3.8-Max-0902, featuring 2.4T parameters, a 1M-token context window, and stronger post-training on coding and collaboration.…
Sayak Paul announces an ECCV 2026 tutorial on post-training approaches and lessons from large models like Flux2/3, with a website to bookmark.
Sam Altman says OpenAI spent the summer sprinting on safety priorities and is close to launching its next model, while caution remains essential as capabilities…
MiniMax H3, combined with vLLM-Omni and FastVideo's FastH3, renders a 10.1-second video with synchronized audio in just 8.7 seconds, making real-time interactiv…
Fable 5.1, described as Claude's most advanced coding model, is now available in Command Code on Max plans and API, with 75% cheaper cache reads.
MiniMax H3 open weights are one month old. A developer built a generative video classroom with fal H3 Max, creating animated explainer videos on demand.
A new ML Advocate Engineer at Hugging Face will focus on showcasing LeRobot across different robotics hardware, starting with Menlo Research's Asimov robot.