#91 Gemma 4 26B A4B
none- Cost
- $0.001
- Time
- 39.5s
- Tokens
- 790 tok
Google
Release: 2026-04-03
Tested on: 2026-05-22 00:29
google/gemma-4-26b-a4b-it::none
25.2B total (3.8B active)MoEOpen source
(medium)
(none)
Summary
Gemma 4 26B A4B scores 6.2 on AI BENCHY and ranks #91. It has 10.0 reliability, a 46.7% pass rate, $0.006 total cost, and 6.00s average response time.
What makes Gemma 4 26B A4B unique: Its total benchmark cost is unusually low for its score range.
Model facts
Researched on 2026-08-12
- Parameters
- 25.2B total (3.8B active)
- Architecture
- MoE
- Availability
- Open source
- License
- Apache-2.0
6.2
Consistency
9.2
10.0
$0.006
Total Output Tokens
1,815
Total Input Tokens
0
Input Price
$0.060 / 1M
Output Price
$0.330 / 1M
Flaky tests
2
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 | 5.6 | 10.0 | $0.023 ↑ | Compare | |
| 2026-07-16 23:02 New test added | 5.5 | 10.0 | $0.018 ↓ | Compare | |
| 2026-06-04 14:04 New test added | 6.0 | 10.0 | $0.004 ↓ | Compare | |
| 2026-05-22 00:29 Suite changed | 6.2 | 10.0 | $0.006 | Current run | |
| 2026-04-22 21:39 First recorded run | 6.2 | N/A | $0.005 | 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 00:29 · Suite changed | 60/60 attempts | 6.2 | 9.2 | 10.0 | 8/20 | 2 | 1,815 | 0 | $0.006 | 6.00s |
| 2026-04-22 21:39 · First recorded run | 54/54 attempts | 6.2 | 8.7 | N/A | 7/18 | 3 | 1,783 | 0 | $0.005 | 6.59s |
| Difference | — | 0.0 | +0.5 | +1 | -1 | +32 | 0 | +$0.001 | -591ms |
Benchmark coverage differs: 60/60 attempts (Target: 3 repeats per test) versus 54/54 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 26B A4BnoneFree AvailablevsKAT-Coder-Air V2.5highGemma 4 26B A4BnoneFree AvailablevsKAT-Coder-Air V2.5mediumGemma 4 26B A4BnoneFree AvailablevsQwen3.5-122B-A10BnoneGemma 4 26B A4BnoneFree AvailablevsNemotron 3.5 LightninghighFree AvailableGemma 4 26B A4BnoneFree AvailablevsKimi K2.6noneGemma 4 26B A4BnoneFree AvailablevsQwen3.6 27BnoneGemma 4 26B A4BnoneFree AvailablevsGLM 5.1noneGemma 4 26B A4BnoneFree AvailablevsQwen3.8 27BnoneGemma 4 26B A4BnoneFree AvailablevsLing-3.0-flashnoneGemma 4 26B A4BnoneFree AvailablevsDeepSeek V4 Flash 0423none
Category Breakdown
| Category | Score | Consistency | Tests Correct |
|---|---|---|---|
| Anti-AI Tricks | 8.3 | 10.0 | |
| Coding | 4.1 | 5.8 | |
| Combined | 3.0 | 10.0 | |
| Data parsing and extraction | 10.0 | 10.0 | |
| Domain specific | 3.6 | 7.2 | |
| General Intelligence | 4.0 | 10.0 | |
| Instructions following | 6.3 | 10.0 | |
| Puzzle Solving | 6.2 | 10.0 | |
| Tool Calling | 10.0 | 10.0 | |
| Trivia | 3.0 | 10.0 |