Alibaba announces Qwen3.8-Max with 2.4T parameters and upcoming open weights

Alibaba has announced Qwen3.8-Max as its new flagship model, describing it as a 2.4T-parameter sparse architecture focused on long-horizon agentic work, coding, and multimodal reasoning. The company confirmed that open weights for Qwen3.8-Max will be released next week alongside the open-weight Qwen3.8-27B variant.

API pricing is set at $2.00 per million input tokens, $6.00 per million output tokens, and $0.25 per million cached tokens.
The model features a 1M-token context window, 95B active parameters per token, and native multimodal planning capabilities.
Third-party evaluations place Qwen3.8-Max at #4 in Frontend Code Arena with 1,668 Elo and #2 in Vision Arena with 1,305 Elo.
Vals AI reported a score of 66.1 on the Vals Index, matching Claude Opus 4.7 while costing approximately 2.3x less per test.

The announcement positions Qwen3.8-Max as a direct competitor to top Western closed models, particularly in coding and agentic workflows, though operational accessibility remains a point of discussion.

Benchmarks

Benchmark	Model	Score
Terminal-Bench 2	Qwen3.8-Max	67.4%

Benchmark

Model

Score

Terminal-Bench 2

Qwen3.8-Max

67.4%

Benchmarks

Qwen3.8-Max matches Kimi K3 and DeepSeek V4 Flash on benchmarks

Qwen 3.8-Max released, ranking #2 in Vision Arena

InternScience releases Agents-A1-4B, a 4B-parameter model for long-horizon agentic tasks

Qwen releases Qwen2.5 models with expanded dataset and enhanced coding/math capabilities

Qwen 3.8 Max scores 1588 on Debate Benchmark, up from 1462 for Qwen 3.7 Max