Qwen 3.8 Max: What Alibaba's 2.4T Open-Weight Model Actually Delivers (2026)
Verdict: Qwen 3.8 Max is Alibaba's most capable model to date and the first Qwen-Max-class model promised to ship open weights — a genuine inflection in the 2026 open-weight race. Its standalone strengths are long-horizon autonomous coding and general agent work, where its own published benchmarks put it ahead of GPT-5.6 Sol on five of seven agent tasks; but on the hardest frontier-coding bars…
Qwen 3.8 Max is Alibaba's most powerful AI model to date, boasting 2.4 trillion total parameters with 95 billion of them actively engaged during token processing. It was officially launched on August 3, 2026, following a paid preview from July 19 through the model ID qwen3.8-max-preview. This marks the first time a Max-tier Qwen model has promised open weights, a significant development in the ongoing open-weight race of 2026.
Alibaba has indicated that the open weights will be released the following week after the August 3 release.
On benchmark tables, Qwen 3.8 Max demonstrates its proficiency in general agent work, outperforming GPT-5.6 Sol on five out of seven agent benchmarks, including CoWorkBench, WorkSpaceBench, JobBench, WideSearch, and instruction-following tasks. However, when it comes to extreme deep reasoning or challenging single-shot coding tasks, Claude Fable 5 still leads by a margin of 10-15 points.
The model's pricing is competitive, with an API rate of $2.00 per million input tokens and $6.00 per million output tokens, significantly lower than Claude Fable 5 and GPT-5.6 Sol. This makes it the most affordable frontier-tier API available today, and potentially the most capable model that can be self-hosted once the weights are released.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
