Qwen2.5 72B Instruct vs Qwen3 235B A22B
Humanity's Last Exam Humanity's Last Exam: Very hard expert questions across many fields, written to stump today’s models
Qwen2.5 72B Instruct3.6
Qwen3 235B A22B4.2
Arena Text · creative writing Arena Text · creative writing: Human pairwise preferences in the stated text vote pool
Qwen2.5 72B Instruct1,254
Qwen3 235B A22B1,366
| Benchmark | Qwen2.5 72B Instruct | Qwen3 235B A22B |
|---|---|---|
| Humanity's Last Exam | 3.6 | 4.2 |
| Arena Text · creative writing | 1,254 | 1,366 |
Specs
| Spec | Qwen2.5 72B Instruct | Qwen3 235B A22B |
|---|---|---|
| Input, per 1M | $1.40 | $0.70 |
| Output, per 1M | $5.60 | $2.80 |
| Speed, tok/s | 24 | — |
| Context | 131K | 33K |
| Weights | Open | Open |
| Released | Sep 2024 | Apr 2025 |
| Memory (4-bit) | 44 GB | 141 GB |
| Licence | qwen | Apache-2.0 |
| Commercial use | Restricted | Allowed |
| See images | — | — |
| Read PDFs | — | Partial |
| Hear audio | — | — |
| Watch video | — | — |
| Call tools | ✓ | ✓ |
| Force a tool call | Partial | ? |
| Return strict JSON | ✓ | Partial |
| Search the web | ? | ? |
| Run code | ? | ? |
| Use a computer | ? | ? |
What breaks when you switch
Qwen2.5 72B Instruct → Qwen3 235B A22B
Context: 131K → 33K
Memory (4-bit): 44 GB → 141 GB
Qwen3 235B A22B → Qwen2.5 72B Instruct
Nothing known.