Views
No views yet
[!NOTE] These NVFP4s are self-quantized from the original weights, not a repack. The importance matrix keeps low-bit quants closer to the full-precision model.
| Property | Value |
|---|---|
| Base model | google/gemma-4-26B-A4B-it |
| Parameters | 25.2B total / 3.8B active per token |
| Layers | 30 |
| Experts | 128 routed (top-8) |
| Sliding window | 1024 tokens |
| Context length | 256K tokens |
| Vocabulary | 262K |
| Modalities | Text, Image |
| Architecture | Mixture-of-Experts, 128 experts (top-8), hybrid sliding-window (1024) and global attention, 16 attention heads over 8 KV heads, Gemma4ForConditionalGeneration |
| This repo | NVFP4 weights |
| Benchmark | Score |
|---|---|
| MMLU Pro | 82.6% |
| AIME 2026 no tools | 88.3% |
| LiveCodeBench v6 | 77.1% |
| Codeforces ELO | 1718 |
| GPQA Diamond | 82.3% |
| Tau2 (average over 3) | 68.2% |
| HLE no tools | 8.7% |
| HLE with search | 17.2% |
| BigBench Extra Hard | 64.8% |
| MMMLU | 86.3% |
| MMMU Pro | 73.8% |
| OmniDocBench 1.5 (average edit distance, lower is better) | 0.149 |
| MATH-Vision | 82.4% |
| MedXPertQA MM | 58.1% |
| MRCR v2 8 needle 128k (average) | 44.1% |
google/gemma-4-26B-A4B-it, not our own measurements. Quantization preserves the large majority of this; Q4_K_M and up stay close to full precision.AtomicChat/gemma-4-26B-A4B-it-NVFP4 and hit Use this model.vllm serve AtomicChat/gemma-4-26B-A4B-it-NVFP4 --max-model-len 8192| Parameter | Value |
|---|---|
| temperature | 1.0 |
| top_p | 0.95 |
| top_k | 64 |
google/gemma-4-26B-A4B-it.google/gemma-4-26B-A4B-it (original weights).llm-compressor over a calibration corpus.