#76 Gemma 4 31B
none- Cost
- $0.001
- Time
- 12.8s
- Tokens
- 795 tok
Google
Release: 2026-04-02
Tested on: 2026-05-22 00:30
google/gemma-4-31b-it::none
30.7BDenseOpen source
(medium)
(none)
Summary
Gemma 4 31B scores 6.7 on AI BENCHY and ranks #76. It has 10.0 reliability, a 50.0% pass rate, $0.003 total cost, and 3.84s average response time.
What makes Gemma 4 31B unique: It stands out most in Domain specific, where it ranks #1, while Combined is its weakest area at #17. Its total benchmark cost is unusually low for its score range. It is notably fast compared with similar models.
Model facts
Researched on 2026-08-12
- Parameters
- 30.7B
- Architecture
- Dense
- Availability
- Open source
- License
- Apache-2.0
6.7
Consistency
10.0
10.0
$0.003
Total Output Tokens
1,398
Total Input Tokens
0
Input Price
$0.120 / 1M
Output Price
$0.370 / 1M
Flaky tests
0
Flaky tests had mixed outcomes across runs (at least one pass and one fail).
Hamster playing table tennis
Prompt: Create a detailed SVG illustration of a hamster playing table tennis.
Run history
| Tested on | Score | Reliability | Tests Correct | Total Cost | Compare |
|---|---|---|---|---|---|
| 2026-08-14 02:51 Re-test | 6.1 | 10.0 | $0.018 ↓ | Compare | |
| 2026-07-16 23:02 New test added | 6.2 | 10.0 | $0.035 ↑ | Compare | |
| 2026-06-04 14:04 New test added | 6.5 | 10.0 | $0.004 ↓ | Compare | |
| 2026-05-22 00:30 Suite changed | 6.7 | 10.0 | $0.003 | Current run | |
| 2026-04-14 00:56 First recorded run | 6.9 | N/A | $0.003 | Compare |
This run used a different benchmark suite. Keep suite changes in mind when reading historical movement.
Run comparison
| Run | Benchmark coverage | Score | Consistency | Reliability | Tests Correct | Flaky tests | Total Output Tokens | Total Input Tokens | Total Cost | Response Time (avg) |
|---|---|---|---|---|---|---|---|---|---|---|
| 2026-05-22 00:30 · Suite changed | 60/60 attempts | 6.7 | 10.0 | 10.0 | 10/20 | 0 | 1,398 | 0 | $0.003 | 3.84s |
| 2026-08-14 02:51 · Re-test | 66/66 attempts | 6.1 | 9.6 | 10.0 | 9/22 | 1 | 13,317 | 125,703 | $0.018 | 5.41s |
| Difference | — | +0.6 | +0.4 | 0.0 | +1 | -1 | -11919 | -125703 | -$0.015 | -1565ms |
Benchmark coverage differs: 60/60 attempts (Target: 3 repeats per test) versus 66/66 attempts (Target: 3 repeats per test). Totals and repeat-sensitive metrics are not directly comparable.
These two runs used different benchmark suites, so the deltas reflect both model changes and suite changes.
Charts
Choose the first model, then click a second model to open a side-by-side page.
Top Models by Score
Score vs Total Cost
Response Time (avg)
Score vs Response Time (avg)
Total Output Tokens
Score vs Total Output Tokens
Quick Compare
Gemma 4 31BnoneFree AvailablevsQwen3.6 FlashnoneGemma 4 31BnoneFree AvailablevsQwen3.5 Plus 2026-04-20noneGemma 4 31BnoneFree AvailablevsTrinity Large ThinkingmediumGemma 4 31BnoneFree AvailablevsNemotron 3 UltranoneFree AvailableGemma 4 31BnoneFree AvailablevsInklinglowGemma 4 31BnoneFree AvailablevsTrinity Large ThinkinglowGemma 4 31BnoneFree AvailablevsQwen3.5-FlashmediumGemma 4 31BnoneFree AvailablevsGemini 2.5 FlashnoneGemma 4 31BnoneFree AvailablevsGemini 3.1 Flash LitenoneGemma 4 31BnoneFree AvailablevsGPT-5.6 Terranone
Category Breakdown
| Category | Score | Consistency | Tests Correct |
|---|---|---|---|
| Anti-AI Tricks | 6.5 | 10.0 | |
| Coding | 6.8 | 10.0 | |
| Combined | 3.0 | 10.0 | |
| Data parsing and extraction | 10.0 | 10.0 | |
| Domain specific | 7.7 | 10.0 | |
| General Intelligence | 10.0 | 10.0 | |
| Instructions following | 6.5 | 10.0 | |
| Puzzle Solving | 6.5 | 10.0 | |
| Tool Calling | 3.0 | 10.0 | |
| Trivia | 3.0 | 10.0 |