Gemma 4 26B A4B
Google · Released 31 Mar 2026 · gemma-4-26b-a4b-it
–
Input price / 1M tokens
–
Output price / 1M tokens
262K
Context window
#54 of 71
–
Output tokens / second
77.1%
LiveCodeBench
#3 of 7
82.3%
GPQA Diamond
#22 of 27
68.2%
τ²-bench
#6 of 6
8.7%
Humanity’s Last Exam (no tools)
#22 of 23
Summary
Gemma 4 26B A4B is an open-weight model from Google, released on 31 Mar 2026.
Its 262K-token context window ranks #54 of 71.
Mixture-of-experts Gemma 4 open model: 25.2B total and 3.8B active parameters, 256K context. It runs almost as fast as a 4B model.
Benchmarks · 9 reported
| Benchmark | Score | Bar | Rank | Source |
|---|---|---|---|---|
| MMLU Proinstruction-tuned | 82.6% | – | Gemma 4 model cardSelf-reported | |
| AIME 2026no tools | 88.3% | – | Gemma 4 model cardSelf-reported | |
| LiveCodeBench | 77.1% | #3 of 7 | Gemma 4 model cardSelf-reported | |
| Codeforces | 1,718 | Elo-style | – | Gemma 4 model cardSelf-reported |
| GPQA Diamond | 82.3% | #22 of 27 | Gemma 4 model cardSelf-reported | |
| τ²-benchaverage over 3 | 68.2% | #6 of 6 | Gemma 4 model cardSelf-reported | |
| Humanity’s Last Exam (no tools)no tools (17.2% with search) | 8.7% | #22 of 23 | Gemma 4 model cardSelf-reported | |
| MMMU-Pro | 73.8% | #6 of 12 | Gemma 4 model cardSelf-reported | |
| MRCR v2 (8-needle)128k average | 44.1% | #12 of 13 | Gemma 4 model cardSelf-reported |
Self-reported means the lab ran the test itself. A rank appears only where several labs report the same version of a benchmark.
Similar models
Models with the most benchmarks in common with Gemma 4 26B A4B, and the closest scores.
LiveCodeBench
#3 of 7Fresh competitive programming problems
GPQA Diamond
#22 of 27Graduate-level science questions
- Gemma 4 26B A4B82.3
- Gemma 4 31B84.3
- Gemini 3.1 Flash-Lite86.9
- Gemini 3 Flash90.4
- Gemini 2.5 Flash-Lite66.7
- Gemini 3.1 Pro94.3
τ²-bench
#6 of 6Tool use with a simulated customer
- Gemma 4 26B A4B68.2
- Gemma 4 31B76.9
- Gemini 3 Flash90.2
- Gemini 3.1 Pro90.8
Humanity’s Last Exam (no tools)
#22 of 23Expert questions, no search or code
Compare Gemma 4 26B A4B side by side
Pre-filled with the two closest models. Swap any of them.
Sources
- license, context, modalities, parameters, knowledge cutoff, benchmarks: ai.google.dev/gemma/docs/core/model_card_4, checked 2 Oct 2026
- release date: ai.google.dev/gemma/docs/releases, checked 2 Oct 2026
- Gemini API availability and id: ai.google.dev/gemini-api/docs/changelog, checked 2 Oct 2026
- Gemini API price (free only): ai.google.dev/gemini-api/docs/pricing, checked 2 Oct 2026
Building on Gemma 4 26B A4B?
Get the architecture right before the bill arrives.
We size caching, routing and fallbacks for your workload on a free call, and tell you where a cheaper model is good enough.