Gemma 4 26B

Google DeepMind · google/gemma-4-26b

ProviderGoogle DeepMind
Familygemma
Typellm-reasoning
Statusactive
Released2026-04-02
Updated2026-04-02

Reported benchmark scores

These scores carry one collection date for the whole card (2026-04) and a source list rather than a source per score, so they cannot be attributed individually. They are shown as reported and are not verified evidence. Benchmark pages carry the reviewed, dated evidence where it exists.
BenchmarkCatalogue standingScore
Aider Polyglot Benchmarkunassessed52.1
Arena Elo — Codingunassessed1290.0
Arena Elo — Mathunassessed1270.0
Arena Elo — Overall (Text)unassessed1280.0
BIG-Bench Hardunassessed60.1
GPQA Diamondunassessed65.2
HumanEvalunassessed82.5
IFEvalunassessed81.5
MATH-500unassessed85.1
MMLU: Astronomyunassessed72.2
MMLU: Biology (subcategory)unassessed80.8
MMLU: Business Ethicsunassessed73.5
MMLU: Chemistry (subcategory)unassessed73.2
MMLU: Clinical Knowledgeunassessed79.2
MMLU: Computer Science (subcategory)unassessed81.2
MMLU: Jurisprudenceunassessed70.2
MMLU: Physics (unresolved key)unassessed77.5
MMLU-Prounassessed72.3
MMLU: Professional Accountingunassessed64.5
MMLU: Professional Lawunassessed69.2
MultiPL-Eunassessed75.4
MultiPL-E: C++unassessed74.8
MultiPL-E: C#unassessed72.5
MultiPL-E: Gounassessed71.5
MultiPL-E: Javaunassessed79.2
MultiPL-E: JavaScriptunassessed77.5
MultiPL-E: Juliaunassessed52.5
MultiPL-E: Kotlin (unconfirmed)unassessed62.8
MultiPL-E: Luaunassessed45.2
MultiPL-E: Perlunassessed40.5
MultiPL-E: PHPunassessed69.8
MultiPL-E: Pythonunassessed82.5
MultiPL-E: Runassessed48.2
MultiPL-E: Rubyunassessed55.5
MultiPL-E: Rustunassessed65.2
MultiPL-E: Scalaunassessed50.2
MultiPL-E: Swiftunassessed55.5
MultiPL-E: TypeScriptunassessed76.8
MuSRunassessed28.7
SWE-bench Verifiedunassessed28.5

Data

This card as JSON · Edit on GitHub