Qwen3.7-Max
Alibaba · Released 21 May 2026 · qwen3.7-max
$2.50
Input price / 1M tokens
#45 of 61, cheapest first
$7.50
Output price / 1M tokens
#35 of 61, cheapest first
1M
Context window
#29 of 71
203
Output tokens / second
#10 of 61
74.5%
Terminal-Bench 2.1
#14 of 17
60.6%
SWE-bench Pro
#8 of 16
92.4%
GPQA Diamond
#9 of 27
41.4%
Humanity’s Last Exam (no tools)
#8 of 23
Summary
Qwen3.7-Max is a proprietary model from Alibaba, released on 21 May 2026.
At $2.50 input and $7.50 output per million tokens, it is #45 of 61 on input price, cheapest first.
Its 1M-token context window ranks #29 of 71.
Measured output speed is 203 tokens per second, #10 of 61.
Largest model of the Qwen3.7 series, aimed at agent work, programming and long autonomous tasks. The qwen3.7-max alias points to the 2026-05-20 text-only snapshot; the 2026-06-08 snapshot adds image understanding. Not open weight.
Benchmarks · 4 reported
| Benchmark | Score | Bar | Rank | Source |
|---|---|---|---|---|
| Terminal-Bench 2.1 | 74.5% | #14 of 17 | Qwen3.8-2.4T-A95B model card (Qwen3.7-Max column)Self-reported | |
| SWE-bench Pro | 60.6% | #8 of 16 | Qwen3.8-2.4T-A95B model card (Qwen3.7-Max column)Self-reported | |
| GPQA Diamond | 92.4% | #9 of 27 | Qwen3.8-2.4T-A95B model card (Qwen3.7-Max column)Self-reported | |
| Humanity’s Last Exam (no tools)no tools | 41.4% | #8 of 23 | Qwen3.8-2.4T-A95B model card (Qwen3.7-Max column)Self-reported |
Self-reported means the lab ran the test itself. A rank appears only where several labs report the same version of a benchmark.
Similar models
Models with the most benchmarks in common with Qwen3.7-Max, and the closest scores.
Terminal-Bench 2.1
#14 of 17Terminal tasks, earlier set
- Qwen3.7-Max74.5
- GLM-5.281
- Qwen3.8-27B73
- Qwen3.8-Max86.6
SWE-bench Pro
#8 of 16Harder multi-file engineering tasks
- Qwen3.7-Max60.6
- GLM-5.262.1
- Qwen3.8-27B61.7
- Qwen3.8-Max67.7
- GPT-5.558.6
- Qwen3.8-Flash62.5
GPQA Diamond
#9 of 27Graduate-level science questions
- Qwen3.7-Max92.4
- GLM-5.291.2
- Qwen3.8-27B89.2
- Qwen3.8-Max92.6
- GPT-5.593.6
- Qwen3.8-Flash91.7
Humanity’s Last Exam (no tools)
#8 of 23Expert questions, no search or code
- Qwen3.7-Max41.4
- GLM-5.240.5
- Qwen3.8-27B30.8
- Qwen3.8-Max43.6
- GPT-5.541.4
- Qwen3.8-Flash35.9
Compare Qwen3.7-Max side by side
Pre-filled with the two closest models. Swap any of them.
Sources
- price, snapshots: alibabacloud.com/help/en/model-studio/model-pricing, checked 2 Oct 2026
- context 1M, legacy status: alibabacloud.com/help/en/model-studio/text-generation-model, checked 2 Oct 2026
- cache prices, max output, modalities: qwencloud.com/models/qwen3.7-max, checked 2 Oct 2026
- release dates: alibabacloud.com/help/en/model-studio/newly-released-models, checked 2 Oct 2026
- output speed: artificialanalysis.ai/models/qwen3-7-max, checked 2 Oct 2026
Building on Qwen3.7-Max?
Get the architecture right before the bill arrives.
We size caching, routing and fallbacks for your workload on a free call, and tell you where a cheaper model is good enough.