Can I run Gemma 4 12B on an RTX 5080 16GB?
Yes: Gemma 4 12B needs about 8.98 GB at Q4_K_M with 32K tokens of context, which fits the RTX 5080 16GB with 7.02 GB to spare. At 32K the RTX 5080 holds up to Q8_0 (14.6 GB), and Q4_K_M runs up to 256K (full) tokens.
Yes Q4_K_M with 32K tokens of context
- Q4_K_M, 32K
- 8.98 GB
- RTX 5080
- 16 GB, 960 GB/s
- To spare
- 7.02 GB
- Tokens/s
- 58–83 tokens/s
At Q4_K_M with 32K tokens of context it writes about 58–83 tokens/s for one request on an RTX 5080 16GB.
Best precision for Gemma 4 12B on an RTX 5080 16GB
The most precise setting that leaves at least 0.5 GB free; one that fits with less is marked tight.
| Context | Best fit | Memory | Free | Tokens/s |
|---|---|---|---|---|
| 8K | Q8_0 | 14.2 GB | 1.83 GB | 37–53 |
| 32K | Q8_0 | 14.6 GB | 1.42 GB | 36–51 |
| 128K | FP8 | 15.5 GB | 545 MB | 34–48 |
| 256K (full) | Q6_K | 15.5 GB | 550 MB | 34–48 |
Gemma 4 12B on the RTX 5080 as the context fills
One request, FP16 KV cache, 0.5 GB plus 10% overhead; a minus sign is memory missing, and tight is less than 0.5 GB free.
| Context | Q4_K_M | Free | Tokens/s | Q8_0 | Free | Tokens/s |
|---|---|---|---|---|---|---|
| 4K | 8.50 GB | 7.50 GB | 61–88 | 14.1 GB | 1.90 GB | 38–53 |
| 8K | 8.57 GB | 7.43 GB | 61–87 | 14.2 GB | 1.83 GB | 37–53 |
| 16K | 8.70 GB | 7.30 GB | 60–86 | 14.3 GB | 1.69 GB | 37–52 |
| 32K | 8.98 GB | 7.02 GB | 58–83 | 14.6 GB | 1.42 GB | 36–51 |
| 64K | 9.53 GB | 6.47 GB | 55–78 | 15.1 GB | 887 MB | 35–49 |
| 128K | 10.6 GB | 5.37 GB | 49–70 | 16.2 GB | −239 MB | — |
| 256K (full) | 12.8 GB | 3.17 GB | 41–58 | 18.4 GB | −2.43 GB | — |
Gemma 4 12B on more than one RTX 5080
| Cards | Best at 32K | Tokens/s | Q4_K_M longest | Q4_K_M tokens/s |
|---|---|---|---|---|
| 2× (32 GB) | BF16 | 37–54 | 256K (full) | 88–142 |
| 4× (64 GB) | BF16 | 65–100 | 256K (full) | 135–240 |
Tensor parallel with the combined bandwidth, as in vLLM; llama.cpp splits layers by default and runs at about one card's speed.
Other options
- The next smaller setting, Q3_K_M, takes 7.55 GB at 32K, 8.45 GB under the RTX 5080; it fits with 0.5 GB to spare up to 256K (full) tokens.
- Smallest setup for Q4_K_M at 32K: RTX 3060 12GB.
Run Gemma 4 12B on the RTX 5080 with llama-server
llama-server -hf unsloth/gemma-4-12b-it-GGUF:Q4_K_M -c 262144 -np 1gemma-4-12b-it-Q4_K_M.gguf, 7.1 GB, from unsloth/
Questions
Can I run Gemma 4 12B on an RTX 5080 16GB?
Yes: Gemma 4 12B needs about 8.98 GB at Q4_K_M with 32K tokens of context, which fits the RTX 5080 16GB with 7.02 GB to spare. At 32K the RTX 5080 holds up to Q8_0 (14.6 GB), and Q4_K_M runs up to 256K (full) tokens.
How fast is Gemma 4 12B on an RTX 5080 16GB?
At Q4_K_M with 32K tokens of context it writes about 58–83 tokens/s for one request on an RTX 5080 16GB.
What does a second RTX 5080 16GB change for Gemma 4 12B?
Two RTX 5080 16GB cards (32 GB in one tensor-parallel group) hold Gemma 4 12B at BF16 with 32K, and Q4_K_M up to 256K (full) tokens, at about 88–142 tokens/s.
Try other settings in the VRAM calculator or the speed calculator. See also Gemma 4 12B VRAM requirements, what LLMs an RTX 5080 16GB can run and every pair, or detect your own GPU. Model data checked .