Model Comparison

DiffusionGemma 26B-A4B vs Gemma 4 31BWhich is better in 2026?

Gemma 4 31B significantly outperforms across most benchmarks.

Verdict: DiffusionGemma 26B-A4B vs Gemma 4 31B — which is better?

DiffusionGemma 26B-A4B (by Google) and Gemma 4 31B (by Google) are two of the AI models people compare most. Here is how they stack up on benchmarks, price and capabilities, and which one to pick in 2026.

DiffusionGemma 26B-A4B outperforms in 0 benchmarks, while Gemma 4 31B is better at 11 benchmarks (AIME 2026, BIG-Bench Extra Hard, GPQA, Humanity's Last Exam, LiveCodeBench v6, MathVision, MedXpertQA, MMLU-Pro, MMMLU, MMMU-Pro, t2-bench). Gemma 4 31B significantly outperforms across most benchmarks.

Choose DiffusionGemma 26B-A4B if…

  • you want the most recent training data — it shipped Jun 2026

Choose Gemma 4 31B if…

  • you want the strongest raw capability — it leads on 11 of 11 shared benchmarks

Performance Benchmarks

Comparative analysis across standard metrics

11 benchmarks

DiffusionGemma 26B-A4B outperforms in 0 benchmarks, while Gemma 4 31B is better at 11 benchmarks (AIME 2026, BIG-Bench Extra Hard, GPQA, Humanity's Last Exam, LiveCodeBench v6, MathVision, MedXpertQA, MMLU-Pro, MMMLU, MMMU-Pro, t2-bench).

Gemma 4 31B significantly outperforms across most benchmarks.

Fri Jul 17 2026 • llm-stats.com

Arena Performance

Human preference votes

Model Size

Parameter count comparison

5.5B diff

Gemma 4 31B has 5.5B more parameters than DiffusionGemma 26B-A4B, making it 21.8% larger.

Google
DiffusionGemma 26B-A4B
25.2Bparameters
Google
Gemma 4 31B
30.7Bparameters
25.2B
DiffusionGemma 26B-A4B
30.7B
Gemma 4 31B

Context Window

Maximum input and output token capacity

Only Gemma 4 31B specifies input context (262,144 tokens). Only Gemma 4 31B specifies output context (131,072 tokens).

Google
DiffusionGemma 26B-A4B
Input- tokens
Output- tokens
Google
Gemma 4 31B
Input262,144 tokens
Output131,072 tokens
Fri Jul 17 2026 • llm-stats.com

Input Capabilities

Supported data types and modalities

Both DiffusionGemma 26B-A4B and Gemma 4 31B support multimodal inputs.

They are both capable of processing various types of data, offering versatility in application.

DiffusionGemma 26B-A4B

Text
Images
Audio
Video

Gemma 4 31B

Text
Images
Audio
Video

License

Usage and distribution terms

Both models are licensed under Apache 2.0.

Both models share the same licensing terms, providing consistent usage rights.

DiffusionGemma 26B-A4B

Apache 2.0

Open weights

Gemma 4 31B

Apache 2.0

Open weights

Release Timeline

When each model was launched

DiffusionGemma 26B-A4B was released on 2026-06-10, while Gemma 4 31B was released on 2026-04-02.

DiffusionGemma 26B-A4B is 2 months newer than Gemma 4 31B.

DiffusionGemma 26B-A4B

Jun 10, 2026

1 months ago

2mo newer
Gemma 4 31B

Apr 2, 2026

3 months ago

Knowledge Cutoff

When training data ends

Both models have the same knowledge cutoff date of 2025-01-01.

They should have similar awareness of historical events and information up to this date.

DiffusionGemma 26B-A4B

Jan 2025

Gemma 4 31B

Jan 2025

Outputs Comparison

Notice missing or incorrect data?Start an Issue discussion→

Key Takeaways

Google

No standout differentiators in the data we have for this pair.

Google
Larger context window (262,144 tokens)
Higher AIME 2026 score (89.2% vs 69.1%)
Higher BIG-Bench Extra Hard score (74.4% vs 47.6%)
Higher GPQA score (84.3% vs 73.2%)
Higher Humanity's Last Exam score (26.5% vs 11.9%)
Higher LiveCodeBench v6 score (80.0% vs 69.1%)
Higher MathVision score (85.6% vs 70.5%)
Higher MedXpertQA score (61.3% vs 49.0%)
Higher MMLU-Pro score (85.2% vs 77.6%)
Higher MMMLU score (88.4% vs 81.5%)
Higher MMMU-Pro score (76.9% vs 54.3%)
Higher t2-bench score (86.4% vs 56.2%)

Detailed Comparison

Interactive Arena

Judge for yourself.

Run your own prompts against DiffusionGemma 26B-A4B and Gemma 4 31B side-by-side, then vote on the output you prefer.

DiffusionGemma 26B-A4B
✓ Preferred
Gemma 4 31B
Open in Playground
AI Model Comparison Table
Feature
Google
DiffusionGemma 26B-A4B
Google
Gemma 4 31B

FAQ

Common questions about DiffusionGemma 26B-A4B vs Gemma 4 31B.

Which is better, DiffusionGemma 26B-A4B or Gemma 4 31B?

Gemma 4 31B significantly outperforms across most benchmarks. DiffusionGemma 26B-A4B is made by Google and Gemma 4 31B is made by Google. The best choice depends on your use case — compare their benchmark scores, pricing, and capabilities above.

How does DiffusionGemma 26B-A4B compare to Gemma 4 31B in benchmarks?

DiffusionGemma 26B-A4B scores MMMLU: 81.5%, MMLU-Pro: 77.6%, GPQA: 73.2%, MathVision: 70.5%, AIME 2026: 69.1%. Gemma 4 31B scores AIME 2026: 89.2%, MMMLU: 88.4%, t2-bench: 86.4%, MathVision: 85.6%, MMLU-Pro: 85.2%.

What are the context window sizes for DiffusionGemma 26B-A4B and Gemma 4 31B?

DiffusionGemma 26B-A4B supports an unknown number of tokens and Gemma 4 31B supports 262K tokens. A larger context window lets you process longer documents, conversations, or codebases in a single request.

Related comparisons