Google model product
DiffusionGemma 26B A4B
DiffusionGemma 26B A4B is Google DeepMind’s experimental open-weights text diffusion model based on the Gemma 4 26B-A4B Mixture-of-Experts architecture.
Updated Sep 8, 2026. Default version: DiffusionGemma 26B-A4B
LLMBoard Score32.5DiffusionGemma 26B-A4B
Coverage40%14 benchmark families
Context windowN/ATokens
Official input priceN/AOfficial price unavailable
DiffusionGemma 26B A4B Capability Profile
This profile uses the model's current scored version. Arena ratings and prices are shown separately.
DiffusionGemma 26B A4B Benchmark Results
Benchmark scores for DiffusionGemma 26B-A4B.
Sort by
Benchmark | Score | Rank | Participants | Percentile | Evidence | Evaluated |
|---|
| BenchmarkMRCR v2 | Score32.00% | Rank02 | Participants3 | Percentile50.00% | EvidenceC | Evaluated |
| BenchmarkBIG-Bench Extra Hard | Score47.60% | Rank04 | Participants11 | Percentile70.00% | EvidenceC | Evaluated |
| BenchmarkMedXpertQA | Score49.00% | Rank07 | Participants12 | Percentile45.45% | EvidenceC | Evaluated |
| BenchmarkOmniDocBench 1.5 | Score31.90% | Rank15 | Participants18 | Percentile17.65% | EvidenceC | Evaluated |
| BenchmarkCodeForces | Score47.63% | Rank17 | Participants17 | Percentile0.00% | EvidenceC | Evaluated |
| BenchmarkMathVision | Score70.50% | Rank19 | Participants35 | Percentile47.06% | EvidenceC | Evaluated |
| BenchmarkAIME 2026 | Score69.10% | Rank20 | Participants22 | Percentile9.52% | EvidenceC | Evaluated |
| Benchmarkt2-bench | Score56.20% | Rank20 | Participants23 | Percentile13.64% | EvidenceC | Evaluated |
| BenchmarkMMMLU | Score81.50% | Rank37 | Participants49 | Percentile25.00% | EvidenceC | Evaluated |
| BenchmarkLiveCodeBench v6 | Score69.10% | Rank39 | Participants62 | Percentile37.70% | EvidenceC | Evaluated |
| BenchmarkMMMU-Pro | Score54.30% | Rank57 | Participants69 | Percentile17.65% | EvidenceC | Evaluated |
| BenchmarkMMLU-Pro | Score77.60% | Rank68 | Participants138 | Percentile51.09% | EvidenceC | Evaluated |
| BenchmarkHumanity's Last Exam | Score11.90% | Rank87 | Participants103 | Percentile15.69% | EvidenceC | Evaluated |
| BenchmarkGPQA | Score73.20% | Rank128 | Participants247 | Percentile48.37% | EvidenceC | Evaluated |
DiffusionGemma 26B A4B Arena Results
Preference and agent-evaluation results for the default version.
No Arena results
The default version does not have a matching Arena result yet.
DiffusionGemma 26B A4B Pricing
Official vendor API pricing appears first, followed by individual provider offers.
- Official API
- N/A
- Official provider
- N/A
- Lowest third-party
- N/A
- Tracked offerings
- 0
No provider prices
The default version has no current input or output token prices.
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
DiffusionGemma 26B A4B Runtime Performance
Provider-specific output speed and catalog latency for DiffusionGemma 26B-A4B. Runtime does not affect the capability score.
No runtime data
No provider-specific speed or latency record is linked to the default version yet.
Output Speed is generated output tokens received per second. Catalog latency is reported separately from observed provider TTFT.
DiffusionGemma 26B A4B Specifications
Technical details for the model's default version.
- Version
- DiffusionGemma 26B-A4B
- Released
- Jun 10, 2026
- Knowledge cutoff
- Jan 1, 2025
- Parameters
- 25.2B
- Context window
- N/A
- Max output
- N/A
- Inputs
- image, text
- Outputs
- text
- Open weights
- Yes
- License
- Apache 2.0
DiffusionGemma 26B A4B Versions
Available versions of this model. The score column identifies the version used in the overall ranking.
Sort by
Version | Released | LLMBoard | Parameters | Context | Max output | Open weights | License |
|---|
| VersionDiffusionGemma 26B-A4B | Released | LLMBoard32.47 | Parameters25.2B | ContextN/A | Max outputN/A | Open weightsYes | LicenseApache 2.0 |
Models similar to DiffusionGemma 26B A4B
Recommendations prioritize the same model type and family, then the closest LLMBoard score.
GO
Google
Gemini 2.0 Flash Thinking
GOGoogle
Gemma 4 12B
GOGoogle
Gemini 2.0 Flash
GOGoogle
Gemini 2.5 Flash
GOGoogle
Gemma 4 E4B
GOGoogle
Gemini 2.0 Flash Lite
What is DiffusionGemma 26B A4B?
Key information about DiffusionGemma 26B A4B and its available data.
DiffusionGemma 26B A4B uses discrete diffusion to denoise 256-token canvases in parallel and supports text and image inputs. 8 billion active parameters, and a 256K context window.
Data as of 2026-09-08.
FAQ
Common questions about DiffusionGemma 26B A4B.
When was DiffusionGemma 26B A4B released?
DiffusionGemma 26B A4B's default version was released on Jun 10, 2026.
How much does DiffusionGemma 26B A4B cost?
No official standard PAYG price is currently available for DiffusionGemma 26B A4B.
Who created DiffusionGemma 26B A4B?
DiffusionGemma 26B A4B was created by Google.
What is the context window for DiffusionGemma 26B A4B?
A context window is not available for the default version.
Is DiffusionGemma 26B A4B open weight?
Yes. The default version is marked as open weight under Apache 2.0.