Google DeepMind·released 2026-04-021 source
Gemma 4 26B A4B benchmark scores: 6 benchmarks tracked. 0% of its scores are independently reproduced — the rest are self-reported or unverified.
Leads on 0 of 6 benchmarks·best result 82.2 on OTIS Mock AIME 2024-2025 (unverified)·0 of 6 independently reproduced
| Benchmark | Score | Evidence status | Measured |
|---|---|---|---|
| OTIS Mock AIME 2024-2025 | 82.2accuracy (%) | unverifiedT2 | 2026-04-02 |
| GPQA diamond | 64.31accuracy (%) | unverifiedT2 | 2026-04-02 |
| DTBench | 58.22accuracy (%) | unverifiedT1 | 2026-04-02https://conceptualreasoning.ai/dtbench |
| WeirdML | 35.17accuracy (%) | unverifiedT1 | 2026-04-02https://htihle.github.io/weirdml.html |
| LMCA | 34.92accuracy (%) | unverifiedT1 | 2026-04-02https://conceptualreasoning.ai/lmca |
| Chess Puzzles | 1.09accuracy (%) | unverifiedT2 | 2026-04-02 |
Claims drawn from cited facts, not live model generation.
This model was developed by Google DeepMind. It reached release in 2026, placing it among the most recent vintage of models from that developer.
3 cited facts
This model is scored on 6 tracked benchmarks, but it holds no current top score on any of them. Its record is unverified: outside evaluation has neither confirmed nor disputed the reported figures.
3 cited facts
With an average SOTA gap of 42.28 points, this model trails the leading score by a wide margin, placing it well behind the front of the pack, and it holds the top score on none of its benchmarks.
2 cited facts
1 facts cross-checked across data sources: 1 single-source.
Source facts, citations, and refresh stamp for this record.
Sources: epoch_benchmarks
Gemma 4 26B A4B is an AI model developed by Google DeepMind, released 2026-04-02. Its benchmark record below tags every score as independently reproduced, self-reported, or unverified.
Gemma 4 26B A4B has recorded scores on 6 benchmarks, each shown with its evidence status.
Gemma 4 26B A4B has recorded scores on 6 benchmarks — GPQA diamond, OTIS Mock AIME 2024-2025, DTBench, WeirdML, LMCA, Chess Puzzles. The full table above shows each score with its evidence status.
0 of 6 recorded scores (0%) are independently reproduced rather than self-reported by the lab.
Gemma 4 26B A4B has 6 tracked claims: 6 unverified.