Claude 3.7 Sonnet vs Qwen 4 Max (Leaked Staging)
A direct, empirical evaluation of Claude 3.7 Sonnet (Anthropic) against Qwen 4 Max (Leaked Staging) (Alibaba Cloud).
Claude 3.7 Sonnet
Developed by Anthropic · Proprietary parameters
Enterprise software engineering, full-repo refactoring, and agentic workflows.
Qwen 4 Max (Leaked Staging)
Developed by Alibaba Cloud · 1.2 Trillion (Sparse MoE) parameters
Repository-scale software engineering outscoring GPT-6 Sol, multilingual enterprise agents across 120 languages, and low-cost 1M-token processing.
| Metric | Claude 3.7 Sonnet | Qwen 4 Max (Leaked Staging) |
|---|---|---|
| SWE-bench Verified | 70.3% | 78.6% |
| MMLU-Pro | 82.1% | 91.4% |
| MATH-500 | 96.2% | 95.6% |
| Context Window | 200,000 tokens | 1,000,000 tokens |
| Input Pricing (per 1M) | $3.00 | $1.20 |
| Output Pricing (per 1M) | $15.00 | $4.80 |