← All benchmarks
Gemma 4 31B (Reasoning)
Google · Released 2026-04-02 · Independent benchmark scores from Artificial Analysis
Intelligence Index
22.2
Output speed
35.1
tokens/sec
Blended price
$0.00
per 1M tokens
Time to first token
0.97s
Model metadata
| Model ID | cd26a386-4873-46ff-b853-d239050025a2 |
|---|---|
| Slug | gemma-4-31b |
| Release date | 2026-04-02 |
| Provider | |
| Provider slug | |
| Output speed | 35.1 tok/s |
| Time to first token | 0.97s |
| Time to first answer | 50.49s |
| Blended price ($/1M) | $0.00 |
| Input price ($/1M) | $0.00 |
| Output price ($/1M) | $0.00 |
All evaluation scores (17)
| aime | — |
|---|---|
| aime 25 | — |
| artificial analysis coding index | 43.40 |
| artificial analysis intelligence index | 22.20 |
| artificial analysis math index | — |
| gpqa | 85.7% |
| hle | 23.6% |
| ifbench | 75.6% |
| lcr | 69.7% |
| livecodebench | — |
| math 500 | — |
| mmlu pro | — |
| scicode | 45.5% |
| tau banking | 14.8% |
| tau2 | 59.9% |
| terminalbench hard | 36.4% |
| terminalbench v2 1 | 43.4% |
Source: artificialanalysis.ai/models/gemma-4-31b
Benchmark data from Artificial Analysis.