Qwen3.6 Plus vs Qwen3.7 Plus
Humanity's Last Exam Humanity's Last Exam: Very hard expert questions across many fields, written to stump today’s models
Qwen3.6 Plus27.8
Qwen3.7 Plus35.6
Terminal-Bench 4.0 Terminal-Bench 4.0 · source setting 2: Agents use a terminal to complete executable coding, system administration and research tasks in the 4.0 suite
Qwen3.6 Plus—
Qwen3.7 Plus1
AutomationBench-AA AutomationBench-AA: Agents orchestrate SaaS workflows through APIs; objectives receive credit only when the task has no guardrail violation
Qwen3.6 Plus—
Qwen3.7 Plus17.4
Arena Text · creative writing Arena Text · creative writing: Human pairwise preferences in the stated text vote pool
Qwen3.6 Plus1,407
Qwen3.7 Plus1,431
MMMU-Pro (vision) MMMU-Pro (vision): College-level questions that require understanding images: charts, diagrams, photos
Qwen3.6 Plus78
Qwen3.7 Plus80.5
| Benchmark | Qwen3.6 Plus | Qwen3.7 Plus |
|---|---|---|
| Humanity's Last Exam | 27.8 | 35.6 |
| Terminal-Bench 4.0 | — | 1 |
| AutomationBench-AA | — | 17.4 |
| Arena Text · creative writing | 1,407 | 1,431 |
| MMMU-Pro (vision) | 78 | 80.5 |
Specs
| Spec | Qwen3.6 Plus | Qwen3.7 Plus |
|---|---|---|
| Input, per 1M | $0.50 | $0.32 |
| Output, per 1M | $3 | $1.28 |
| Cost per task? | — | $0.22 |
| Speed, tok/s | 56 | 58 |
| Context | 1M | 1M |
| Weights | Closed | Closed |
| Released | Apr 2026 | Jun 2026 |
| See images | ✓ | ✓ |
| Read PDFs | Partial | Partial |
| Hear audio | — | — |
| Watch video | ✓ | — |
| Call tools | ✓ | ✓ |
| Force a tool call | Partial | Partial |
| Return strict JSON | ✓ | ✓ |
| Search the web | ? | ? |
| Run code | ? | ? |
| Use a computer | ? | ? |
What breaks when you switch
Qwen3.6 Plus → Qwen3.7 Plus
Nothing known.
Qwen3.7 Plus → Qwen3.6 Plus
Nothing known.