Google logo

Gemma 4 26B-A4B

Multimodal
Google

Gemma 4 26B-A4B is Google's open-weight mixture-of-experts model with 25.2 billion total parameters (4B active), released under the Apache 2.0 license. It delivers strong multilingual, mathematical, and multimodal reasoning for its size, with a 262K token context window.

Key Specifications

Parameters
25.2B
Context
262.1K
Release Date
April 2, 2026
Average Score
70.2%

Timeline

Key dates in the model's history
Announcement
April 2, 2026
Last Update
August 27, 2026
Today
October 17, 2026

Technical Specifications

Parameters
25.2B
Training Tokens
-
Knowledge Cutoff
January 31, 2025
Family
-
Capabilities
MultimodalZeroEval

Pricing & Availability

Input (per 1M tokens)
$0.13
Output (per 1M tokens)
$0.40
Max Input Tokens
262.1K
Max Output Tokens
131.1K
Supported Features
Function CallingStructured OutputCode ExecutionWeb SearchBatch InferenceFine-tuning

Benchmark Results

Model performance metrics across various tests and benchmarks

Reasoning

Logical reasoning and analysis
GPQA
• Self-reported
82.3%

Other Tests

Specialized benchmarks
AIME 2026
No tools • Self-reported
88.0%
MMMLU
• Self-reported
86.0%
t2-bench
Retail • Self-reported
85.5%
MMLU-Pro
• Self-reported
83.0%
BIG-Bench Extra Hard
• Self-reported
64.8%
Humanity's Last Exam
With search • Self-reported
17.2%
LiveCodeBench v6
• Self-reported
77.1%
MathVision
• Self-reported
82.4%
MedXpertQA
MM • Self-reported
58.1%
MMMU-Pro
• Self-reported
73.8%
MRCR v2 (8-needle)
128k • Self-reported
44.1%

License & Metadata

License
apache_2_0
Announcement Date
April 2, 2026
Last Updated
August 27, 2026

Similar Models

All Models

Recommendations are based on similarity of characteristics: developer organization, multimodality, parameter size, and benchmark performance. Choose a model to compare or go to the full catalog to browse all available AI models.