4 min read
Qwen3.8-Max is a 2.4T MoE that codes…
Alibaba's Qwen3.8-Max packs 95B active parameters into a 2.4T MoE stack, ranks ahead of Claude Fable 5 on WebDev Arena, and prices at $2/$6 per million tokens. Weights hit Hugging Face next week.
4 min read
Qwen-AgentWorld simulates seven agent environments in one…
Qwen-AgentWorld is a 35B Apache 2.0 world model that predicts terminal, web, and MCP responses so you can train agents without spinning real sandboxes. Sim RL beat real RL on live search.
4 min read
Qwen-VLA: one vision-language-action model for 11 robot…
Alibaba's Tongyi Lab shipped Qwen-VLA, a unified policy that handles manipulation, navigation, and trajectory prediction across 11 robot embodiments via prompt conditioning. Here is why that pattern matters beyond robotics.
3 min read
Qwen3.7 Max packs 1M tokens and 35-hour…
Alibaba's Qwen3.7 Max brings a 1 million token context window, strong SWE-Bench scores, and Anthropic-compatible APIs to agent workloads. Here is what that means if you ship coding agents for a living.