Written by Jakub Rusinowski · Last updated March 31, 2026
Model library → Gemma 4 → Gemma 4 31B
Gemma 4's flagship open-weight model. Full dense 31B with text + image + audio + video multimodal capability — the only sub-50B open model with native video understanding. Scores 92% on MMLU and 86.4% on HumanEval, rivalling GPT-4o on most benchmarks. Needs ~18–20 GB at Q4_K_M (fits on RTX 3090/4090 or 24 GB Mac). Apache 2.0 licensed.
Gemma 4 31B needs about 20 GB of VRAM at Q4_K_M — quantized weights plus framework overhead, before any KV cache. On Apple Silicon that figure comes out of unified memory.
| Parameters | 31 Billion |
| Context window | 256,000 |
| Architecture | Dense Transformer + Full Multimodal Encoder (image + audio + video) |
| Provider | |
| Licence | Apache 2.0 |
| Specified at | Q4_K_M |
| System RAM | 40 GB |
| Record updated | 2026-03-31 |
Apache-2.0 — commercial use permitted. Commercial use permitted. No usage restrictions beyond attribution.
Modelled on a reference NVIDIA RTX 4090 (24 GB), with no KV cache (this record has no published architecture). Speed figures are ESTIMATES from the memory-bandwidth roofline described on the methodology page, not benchmarks we ran — rows marked measured come from published or reader-submitted runs. VRAM here includes the KV cache, so it reads higher than the headline figure above, which does not.
| Quant | Weights | VRAM needed | Est. speed | Fit on 24 GB |
|---|---|---|---|---|
| Q2_K | 10.2 GB | 11.0 GB | ~59 tok/s (est.) | Fits comfortably |
| Q3_K_M | 13.2 GB | 14.0 GB | ~48 tok/s (est.) | Fits comfortably |
| Q4_K_M | 18.7 GB | 19.5 GB | ~35 tok/s (est.) | Fits comfortably |
| Q5_K_M | 22.0 GB | 22.8 GB | ~31 tok/s (est.) | Tight fit |
| Q6_K | 25.4 GB | 26.2 GB | ~4 tok/s (est.) | Offloads to system RAM (slow) |
| Q8_0 | 32.9 GB | 33.7 GB | ~3 tok/s (est.) | Offloads to system RAM (slow) |
| F16 | 62.0 GB | 62.8 GB | — | Won't fit |
Want the memory numbers alone, at every quantization level and your own context length? Use the Gemma 4 31B VRAM calculator.
or compare on Vast.ai from $0.35/hr (typical low · varies)
As an Amazon Associate we earn from qualifying purchases. Cloud GPU links are referral links — we may earn a commission at no extra cost to you.
The cheapest catalogued GPU that runs Gemma 4 31B is the AMD Radeon RX 7900 XT (20 GB).
Install Ollama, then run:
ollama run gemma4:31bWeights on Hugging Face: google/gemma-4-31B-it.
Quality scores as published by the model's authors or an independent evaluator — not throughput, and not measured by us.
| Benchmark | Score | Provenance |
|---|---|---|
| MMLU | 92.1 / 100 % | reported |
| HumanEval | 86.4 / 100 % | reported |
| MATH | 80.3 / 100 % | reported |
| MMMU (Multimodal) | 74.8 / 100 % | reported |
Best for: reasoning, full multimodal, video analysis, coding, long context, near frontier.
← All Gemma 4 models | VRAM calculator | Build a PC for this model | Check your own hardware