Qwen 3.8 Max: 2.4T-parameter multimodal model, open-weight expected
Prompt Engineering · youtube · 2026-08-03
Qwen 3.8 Max has been released, a multimodal model with 2.4T total parameters and 95B active, expected to go open-weight soon. It ranks top-five on Arena and shows strong coding and agentic performance. The model demonstrates long-horizon agentic capabilities, including autonomously creating and evolving its own harness ("Oh My CLI") and autonomous chip design. The author ran hands-on tests: ISS tracker, Pokémon encyclopedia, crowd animation (where it "cheats" by overlaying text), and a 3D map of LA. API pricing is compared with Kimi K3, and open-weight licensing uncertainty and enterprise vs. general-public tradeoffs are discussed.
Related event: Alibaba releases Qwen3.8-Max, open-sources weights next week(44 posts)→
More from Models
- Ethan Mollick: benchmarks miss long-horizon work; small models need orchestration — emollick · 2026-10-16
- Users report Gemini's memories can't actually be deleted even after wiping chat history — ThrowRA_Purpose2025 · 2026-10-16
- OpenAI stealth-testing mysterious internal model 'Aeon' in Codex desktop app and Dots — cedric_chee · 2026-10-16
- Argon edges out astra as daily driver, but Claude still wins heavy coding work — iruletheworldmo · 2026-10-16
- NVIDIA ships FP4-quantized Qwen3.8-Flash for local use, trending on Hugging Face — nvidia · 2026-10-16
- RoboVista benchmark: Gemini-4 Argon hits 81.9% on robotics understanding — berkeley_ai · 2026-10-16