Qwen3.8-Max-0902 leads on the shared-benchmark average
Ahead on 3 of 4 shared benchmarks, averaging 7.0 points higher
Summarised only from the 4 percentage-scale benchmarks scored by every selected model; details are below.
“Best available” takes each model’s highest recorded non-parallel mode per benchmark, so it may combine modes into a virtual configuration that does not exist. Read it with the mode breakdown.
DeepSeek-V4-Pro
DeepSeek-AI
Benchmark-by-benchmark comparison. Changing the thinking mode or tool filters updates the chart and table below.
“Best available” picks the highest non-parallel mode separately for each benchmark. The resulting series can combine several reasoning levels and is not one reproducible runtime configuration. Choose a mode filter to compare like-for-like runs.
Data sourced primarily from official releases (GitHub, Hugging Face, papers), then benchmark leaderboards, then third-party evaluators. Learn about our data methodology
Every model and runtime mode, benchmark by benchmark. Values are comparable along a row, not between different benchmarks.
4 benchmarks with comparable scores. Each model shows its best score; mode label is displayed below.
| Benchmark | DeepSeek-V4-Pro | Qwen3.8-Max-0902 |
|---|---|---|
DeepSWE 编程与软件工程 | 62.70Thinking Level · Extra High ｜ Tools | 69.30Thinking Level · Extra High ｜ Tools |
NL2Repo-Bench 编程与软件工程 | 61.50Thinking Level · Extra High ｜ Tools | 64.90Thinking Level · Extra High ｜ Tools |
AutomationBench AI Agent - 工具使用 | 31.80Thinking Level · Extra High ｜ Tools | 50.80Thinking Level · Extra High ｜ Tools |
Toolathlon-Verified AI Agent - 工具使用 | 74.10Thinking Level · Extra High ｜ Tools | 73.30Thinking Level · Extra High ｜ Tools |
Official list prices per model API, split by input and output. Unit: USD per 1M tokens.
Architecture, licensing and API modalities. "Not provided" means the field is missing from our database.
| Features & specs | DeepSeek-V4-ProDeepSeek-AI | Qwen3.8-Max-0902阿里巴巴 |
|---|---|---|
Core specsRelease | 2026-08-13 | 2026-09-02 |
Context length | 1M | 1M |
Total parameters | 1.6T | 2.4T |
Active parameters | 49B | 95B |
Max output length | 384,000 tokens | 131,072 tokens |
Architecture | MoE (mixture of experts) | MoE (mixture of experts) |
Runtime modes | 关闭高最高 | 低高极高 |
Availability & licensingCode availability | Available · MIT License | Not public |
Weight availability | Available · MIT License | Not public |
Use & commercial terms | 免费商用授权 | Official service only; subject to provider terms |
Local deploymentWeights | Hugging Face | Not provided |
API modality supportText Input/Output | Input：YesOutput：Yes | Input：YesOutput：Yes |
Image Input/Output | Input：NoOutput：No | Input：YesOutput：No |
Video Input/Output | Input：NoOutput：No | Input：YesOutput：No |
ResourcesPaper / report | DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence | Qwen3.8-Max just got upgraded. Meet Qwen3.8-Max-0902! |
Qwen3.8-Max-0902
阿里巴巴
