Mistral on NVIDIA Blackwell

Can I run Mistral NeMo 12B on an RTX 5070 Ti?

Yes. Mistral NeMo 12B at Q4_K_M uses 9.1 GB of the 14.2 GB available on a RTX 5070 Ti, and runs at about 76.1 tokens per second. You can push the context to 41K.

Runs comfortably

9.1 GB of 14.2 GB · 64%
014 GB
Weights 7.0 GB
KV cache 1.3 GB
Runtime overhead 0.8 GB

Mistral NeMo 12B at Q4_K_M leaves 5.1 GB spare on a RTX 5070 Ti. There is room to raise the context length or move up a quantisation level.

Generation76.1tok/s
Prompt processing1509tok/s
Max context41Ktokens
KV per 1K tokens0GB

Every quantisation of Mistral NeMo 12B on a RTX 5070 Ti

Highlighted row is the highest quality that still fits at 8K context.

QuantisationbpwWeights TotalVerdictMax ctxtok/s
FP16 / BF16 16.00 22.8 GB 24.9 GB Won't fit — 3.51
INT8 / W8A8 8.50 11.8 GB 13.9 GB Fits, but tight 10K 47.1
Q8_0 (GGUF) 8.50 11.8 GB 13.9 GB Fits, but tight 10K 47.1
FP8 (E4M3) 8.00 11.4 GB 13.5 GB Fits, but tight 13K 48.6
Q6_K 6.56 9.4 GB 11.4 GB Runs comfortably 26K 58.5
Q5_K_M 5.67 8.1 GB 10.2 GB Runs comfortably 34K 66.9
AWQ 4-bit 4.25 7.9 GB 10.0 GB Runs comfortably 35K 68.4
GPTQ 4-bit 4.25 7.9 GB 10.0 GB Runs comfortably 35K 68.4
MXFP4 4.25 7.9 GB 10.0 GB Runs comfortably 35K 68.4
Q5_K_S 5.52 7.9 GB 10.0 GB Runs comfortably 35K 68.6
Q4_K_M 4.85 7.0 GB 9.1 GB Runs comfortably 41K 76.1
Q4_K_S 4.58 6.7 GB 8.8 GB Runs comfortably 43K 79.6
Q4_0 4.55 6.6 GB 8.7 GB Runs comfortably 43K 80.0
IQ4_XS 4.25 6.3 GB 8.3 GB Runs comfortably 46K 84.3
Q3_K_M 3.91 5.8 GB 7.9 GB Runs comfortably 48K 89.8
IQ3_M 3.70 5.6 GB 7.6 GB Runs comfortably 50K 93.6
IQ3_XXS 3.06 4.7 GB 6.8 GB Runs comfortably 55K 107
Q2_K 2.63 4.2 GB 6.3 GB Runs comfortably 59K 119
IQ2_XXS 2.06 3.5 GB 5.6 GB Runs comfortably 64K 140
IQ1_M 1.75 3.1 GB 5.2 GB Runs comfortably 66K 154

Also worth checking

Mistral NeMo 12B on every GPU
Full hardware tableEverything that runs on a RTX 5070 Ti
Full model tableWhy the KV cache is the part that bites
Guide