Google DiffusionGemma 26B A4B: discrete diffusion model denoising 256-token canvases in parallel. Gemma 4 MoE with bidirectional attention and
Capabilities, design details, and architectural traits
DiffusionGemma replaces token-by-token autoregression with discrete text diffusion. Instead of predicting one token at a time, it starts with a canvas of random placeholder tokens and iteratively denoises them in parallel, shifting the inference bottleneck from memory bandwidth to compute.
| Trait | Detail |
|---|---|
| Generation method | Block-autoregressive multi-canvas sampling: a 256-token canvas is denoised in parallel via a diffusion sampler, committed to the KV cache, then the next canvas begins |
| Architecture | Encoder-decoder design: an autoregressive encoder processes and caches the prompt, while the decoder applies bidirectional attention over the generation canvas via cross-attention |
| Self-correction | Bidirectional context lets every canvas position attend to all others simultaneously, enabling real-time error correction and re-noising when confidence drops |
| Speed profile | Optimized for small-batch inference; parallel denoising of 256 tokens yields 15-20 tokens per forward pass |
| MoE design | 8 active experts out of 128 total plus 1 shared, for low-memory local execution |
The model uses Uniform State Diffusion: over multiple denoising passes, highly confident tokens help resolve adjacent positions, causing the full sequence to converge. For outputs longer than 256 tokens, the block-autoregressive mechanism commits each finished canvas to the KV cache before initializing a fresh one conditioned on prior history, combining parallel block speed with sequential stability.
Independent evaluations · Artificial Analysis
Evaluate specifications, pricing, and independent benchmark indices
| Model Details | |||
|---|---|---|---|
| General Info | |||
| Provider | Anthropic | Anthropic | |
| Release Date | June 10, 2026 | July 24, 2026 | June 9, 2026 |
| Knowledge Cutoff | - | May 2026 | - |
| Context & Limits | |||
| Context Window | 256K | 1M Best Context Window | 1M Best Context Window |
| Pricing (per 1M tokens) | |||
| Input Pricing | Free Best Input Pricing | $5 | $10 |
| Output Pricing | Free Best Output Pricing | $25 | $50 |
| Modalities | |||
| Inputs | textimagevideo | textimage | textimagefile |
| Outputs | text | text | text |
| Benchmarks (0-100) | |||
| Intelligence Index | 13.5 | 63.1 Best Intelligence Index | 62.1 |
| Coding Index | 19.7 | 78.0 Best Coding Index | 76.5 |
| Agentic Index | 2.2 | 59.2 Best Agentic Index | 56.6 |
Graduate-level reasoning and expert Q&A evaluation.
Extremely difficult logical reasoning and knowledge.
Logical reasoning over long context windows.
Independent evaluation data provided by Artificial Analysis. To view the latest benchmarks and full details, visit their official site.
