Home / Models / Gemma 4 31B Instruct
Gemma 4 31B Instruct
31B parameters · native context 262,144 · gemma4
Estimated VRAM requirement
Weight memory uses parameter count × bits per weight, plus a quantization overhead factor. KV cache uses standard GQA formulas. Values are calculated estimates, not measured allocations.
| Quantization | Estimated weights | 2K total | 4K total | 8K total | 16K total | 32K total | 64K total | 128K total | 256K total |
|---|---|---|---|---|---|---|---|---|---|
| FP16 | 59.4 GB | 66.8 GB | 68.7 GB | 72.4 GB | 79.9 GB | 94.9 GB | 124.9 GB | 184.9 GB | 304.9 GB |
| Q8 | 33.4 GB | 38.7 GB | 40.6 GB | 44.3 GB | 51.8 GB | 66.8 GB | 96.8 GB | 156.8 GB | 276.8 GB |
| Q6 | 26.0 GB | 30.7 GB | 32.6 GB | 36.4 GB | 43.9 GB | 58.9 GB | 88.9 GB | 148.9 GB | 268.9 GB |
| Q5 | 22.0 GB | 26.4 GB | 28.3 GB | 32.0 GB | 39.5 GB | 54.5 GB | 84.5 GB | 144.5 GB | 264.5 GB |
| Q4 | 18.3 GB | 22.4 GB | 24.3 GB | 28.1 GB | 35.6 GB | 50.6 GB | 80.6 GB | 140.6 GB | 260.6 GB |
| Q3 | 14.7 GB | 18.5 GB | 20.3 GB | 24.1 GB | 31.6 GB | 46.6 GB | 76.6 GB | 136.6 GB | 256.6 GB |
Find a GPU for this model
Run at the quantization and context selected below. Defaults are Q4 and 8K when you do not change the controls. Results are full-VRAM fits only — not a speed ranking.
- Run at
- Q4 · 8K context
- Required VRAM
- 28.1 GB
Current cheapest full-VRAM option at Q4 · 8K: NVIDIA RTX 5000 32 GB $4479.96. See all value options · Build a PC for this model · Compatible workstations · Rent a server · Buy vs rent.
Lowest-cost full-VRAM option
Sorted by lowest current Amazon price among GPUs that fit entirely in VRAM. This is not a performance ranking.
NVIDIA RTX 5000 32 GB
- GPU VRAM
- 32 GB advertised
- Required VRAM
- 28.1 GB
- Usable GPU VRAM
- 28.8 GB
- VRAM headroom
- 0.7 GB — leftover usable memory after the calculated requirement, not a speed ranking
- Compatible
- Yes — fits entirely in VRAM
- Product SKUs
- 2 mapped · 1 Amazon-matched
- Amazon
- $4479.96 current Amazon price
NVIDIA RTX 5090 32 GB
- GPU VRAM
- 32 GB advertised
- Required VRAM
- 28.1 GB
- Usable GPU VRAM
- 28.8 GB
- VRAM headroom
- 0.7 GB — leftover usable memory after the calculated requirement, not a speed ranking
- Compatible
- Yes — fits entirely in VRAM
- Product SKUs
- 8 mapped · 8 Amazon-matched
- Amazon
- $4699.99 current Amazon price
NVIDIA RTX 6000 48 GB
- GPU VRAM
- 48 GB advertised
- Required VRAM
- 28.1 GB
- Usable GPU VRAM
- 43.2 GB
- VRAM headroom
- 15.1 GB — leftover usable memory after the calculated requirement, not a speed ranking
- Compatible
- Yes — fits entirely in VRAM
- Product SKUs
- 1 mapped · 1 Amazon-matched
- Amazon
- $20888.00 current Amazon price
More VRAM headroom
Sorted by leftover usable VRAM after the calculated requirement. Headroom is not tokens/sec.
NVIDIA B200 192 GB
- GPU VRAM
- 192 GB advertised
- Required VRAM
- 28.1 GB
- Usable GPU VRAM
- 172.8 GB
- VRAM headroom
- 144.7 GB — leftover usable memory after the calculated requirement, not a speed ranking
- Compatible
- Yes — fits entirely in VRAM
- Product SKUs
- 0 mapped · 0 Amazon-matched
- Amazon
- Check availability
NVIDIA H200 141 GB
- GPU VRAM
- 141 GB advertised
- Required VRAM
- 28.1 GB
- Usable GPU VRAM
- 126.9 GB
- VRAM headroom
- 98.8 GB — leftover usable memory after the calculated requirement, not a speed ranking
- Compatible
- Yes — fits entirely in VRAM
- Product SKUs
- 0 mapped · 0 Amazon-matched
- Amazon
- Check availability
NVIDIA GH200 96 GB
- GPU VRAM
- 96 GB advertised
- Required VRAM
- 28.1 GB
- Usable GPU VRAM
- 86.4 GB
- VRAM headroom
- 58.3 GB — leftover usable memory after the calculated requirement, not a speed ranking
- Compatible
- Yes — fits entirely in VRAM
- Product SKUs
- 0 mapped · 0 Amazon-matched
- Amazon
- Check availability
Other compatible GPUs
NVIDIA Tesla M10 32 GB
- GPU VRAM
- 32 GB advertised
- Required VRAM
- 28.1 GB
- Usable GPU VRAM
- 28.8 GB
- VRAM headroom
- 0.7 GB — leftover usable memory after the calculated requirement, not a speed ranking
- Compatible
- Yes — fits entirely in VRAM
- Product SKUs
- 1 mapped · 0 Amazon-matched
- Amazon
- Check availability
NVIDIA Quadro GV100 32 GB
- GPU VRAM
- 32 GB advertised
- Required VRAM
- 28.1 GB
- Usable GPU VRAM
- 28.8 GB
- VRAM headroom
- 0.7 GB — leftover usable memory after the calculated requirement, not a speed ranking
- Compatible
- Yes — fits entirely in VRAM
- Product SKUs
- 1 mapped · 0 Amazon-matched
- Amazon
- Check availability
NVIDIA RTX A6000 48 GB
- GPU VRAM
- 48 GB advertised
- Required VRAM
- 28.1 GB
- Usable GPU VRAM
- 43.2 GB
- VRAM headroom
- 15.1 GB — leftover usable memory after the calculated requirement, not a speed ranking
- Compatible
- Yes — fits entirely in VRAM
- Product SKUs
- 2 mapped · 0 Amazon-matched
- Amazon
- Check availability
Compatible GPUs
Status is a calculated memory fit against usable VRAM (advertised × 0.9). Open a GPU to see product SKUs and Amazon listings.
Compare GPUs for this model
Published chip comparisons where both GPUs fully fit this model at Q4 / 8K. Opens with this model preselected; that query is not indexed.
Sources
Model specifications: official repository google/gemma-4-31B-it · revision main. Compatibility: RigForAI calculator v1.0.0 (estimated VRAM). GPU products: Icecat. Amazon is a separate affiliate match and is not used as a spec source.