← All benchmarks
Qwen3.5 4B (Reasoning)
Alibaba · Released 2026-03-02 · Independent benchmark scores from Artificial Analysis
Intelligence Index
13.1
Output speed
16.6
tokens/sec
Blended price
$0.06
per 1M tokens
Time to first token
0.66s
Model metadata
| Model ID | 651ef7ae-9a8f-477e-9c8e-460aa156ba02 |
|---|---|
| Slug | qwen3-5-4b |
| Release date | 2026-03-02 |
| Provider | Alibaba |
| Provider slug | alibaba |
| Output speed | 16.6 tok/s |
| Time to first token | 0.66s |
| Time to first answer | 121.45s |
| Blended price ($/1M) | $0.06 |
| Input price ($/1M) | $0.03 |
| Output price ($/1M) | $0.15 |
All evaluation scores (17)
| aime | — |
|---|---|
| aime 25 | — |
| artificial analysis coding index | 22.60 |
| artificial analysis intelligence index | 13.10 |
| artificial analysis math index | — |
| gpqa | 77.1% |
| hle | 9.9% |
| ifbench | 52.0% |
| lcr | 63.0% |
| livecodebench | — |
| math 500 | — |
| mmlu pro | — |
| scicode | — |
| tau banking | 6.8% |
| tau2 | 92.1% |
| terminalbench hard | 18.2% |
| terminalbench v2 1 | 25.8% |
Source: artificialanalysis.ai/models/qwen3-5-4b
Benchmark data from Artificial Analysis.