Google model product
Gemini Diffusion is a state-of-the-art, experimental text diffusion model from Google DeepMind.
Updated Aug 12, 2026. Default version: Gemini Diffusion
Technical details for the model's default version.
This profile uses the model's current scored version. Arena ratings and prices are shown separately.
Benchmark scores for Gemini Diffusion.
| BigCodeBench | 45.4% | 01 | 2 | 100.0% | C | |
| LBPP (v2) | 56.8% | 01 | 1 | 100.0% | C | |
| Global-MMLU-Lite | 69.1% | 08 | 14 | 46.1% | C | |
| BIG-Bench Extra Hard | 15.0% | 09 | 11 | 20.0% | C | |
| HumanEval | 89.6% | 15 | 66 | 78.5% | C | |
| MBPP | 76.0% | 17 | 33 | 50.0% | C | |
| LiveCodeBench | 30.9% | 61 | 73 | 16.7% | C | |
| SWE-Bench Verified | 22.9% | 103 | 105 | 1.9% | C | |
| AIME 2025 | 23.3% | 110 | 114 | 3.5% | C | |
| GPQA | 40.4% | 205 | 234 | 12.4% | C |
Preference and agent-evaluation results for the default version.
The default version does not have a matching Arena result yet.
Provider-specific output speed and catalog latency for Gemini Diffusion. Runtime does not affect the capability score.
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.
Official vendor API pricing appears first, followed by individual provider offers.
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.
Available versions of this model. The score column identifies the version used in the overall ranking.
| Gemini Diffusion | 3.6 | N/A | N/A | N/A | No | Proprietary |
Key information about Gemini Diffusion and its available data.
Gemini Diffusion is a state-of-the-art, experimental text diffusion model from Google DeepMind. It explores a new kind of language model designed to provide users with greater control, creativity, and speed in text generation.
Instead of predicting text token-by-token, it learns to generate outputs by refining noise step-by-step, allowing for rapid iteration and error correction during generation.
Key capabilities include rapid response times (reportedly 1479 tokens/sec excluding overhead), generation of more coherent text by outputting entire blocks of tokens at once, and iterative refinement for consistent outputs. It excels at tasks like editing, including in math and code contexts.
Data as of 2026-08-11.
Open a comparison with the three ranked models immediately above and below this model.
Recommendations prioritize the same model type and family, then the closest LLMBoard score.
Common questions about Gemini Diffusion.
Gemini Diffusion's default version was released on May 20, 2025.
No official standard PAYG price is currently available for Gemini Diffusion.
Gemini Diffusion was created by Google.
A context window is not available for the default version.
No. The default version is not marked as having publicly available weights.
No provider offering is currently linked to the default version.
Nearby ranked alternatives include GPT-4.1-nano, Nova Micro, Granite 3.3 8B.