Claude Fable 5.5 (Leaked Checkpoint) vs Qwen 2.5 Max
A direct, empirical evaluation of Claude Fable 5.5 (Leaked Checkpoint) (Anthropic) against Qwen 2.5 Max (Alibaba Cloud).
Claude Fable 5.5 (Leaked Checkpoint)
Developed by Anthropic · Proprietary Frontier Dynamic Reasoner parameters
Terminal-Bench 4.0 autonomous shell execution (78.5% target), recursive self-critique, and 1.5M-token repository refactoring.
Qwen 2.5 Max
Developed by Alibaba Cloud · Proprietary (~700B) parameters
Cross-lingual document extraction and structured structured output generation.
| Metric | Claude Fable 5.5 (Leaked Checkpoint) | Qwen 2.5 Max |
|---|---|---|
| SWE-bench Verified | 84.2% | 51.4% |
| MMLU-Pro | 94.6% | 77.2% |
| MATH-500 | 98.4% | 89.4% |
| Context Window | 1,500,000 tokens | 128,000 tokens |
| Input Pricing (per 1M) | $3.50 | $1.60 |
| Output Pricing (per 1M) | $17.50 | $6.40 |