Mistral on NVIDIA Blackwell

Can I run Mistral Small 24B on an RTX 5060 Ti 16GB?

Not at Q4_K_M — it needs 15.5 GB against 14.2 GB available. Drop to IQ4_XS and it fits, at about 23.6 tokens per second.

Won't fit

15.5 GB of 14.2 GB · 109%
016 GB
Weights 13.4 GB
KV cache 1.3 GB
Runtime overhead 0.8 GB
Over the limit 1.3 GB

Short by 1.3 GB. You can run it with 36 of 40 layers on the RTX 5060 Ti 16GB and the rest in system RAM, at roughly 13.1 tok/s — usable for batch work, painful for chat. A smaller quantisation or a shorter context is usually the better trade.

Generation13.1tok/s
Prompt processing419tok/s
Max context0tokens
KV per 1K tokens0GB

Every quantisation of Mistral Small 24B on a RTX 5060 Ti 16GB

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

QuantisationbpwWeights TotalVerdictMax ctxtok/s
FP16 / BF16 16.00 43.9 GB 46.0 GB Won't fit 1.24
INT8 / W8A8 8.50 23.0 GB 25.1 GB Won't fit 3.24
Q8_0 (GGUF) 8.50 23.0 GB 25.1 GB Won't fit 3.24
FP8 (E4M3) 8.00 22.0 GB 24.0 GB Won't fit 3.53
Q6_K 6.56 18.0 GB 20.1 GB Won't fit 5.37
Q5_K_M 5.67 15.6 GB 17.6 GB Won't fit 7.76
Q5_K_S 5.52 15.1 GB 17.2 GB Won't fit 8.50
AWQ 4-bit 4.25 13.5 GB 15.6 GB Won't fit 13.0
GPTQ 4-bit 4.25 13.5 GB 15.6 GB Won't fit 13.0
MXFP4 4.25 13.5 GB 15.6 GB Won't fit 13.0
Q4_K_M 4.85 13.4 GB 15.5 GB Won't fit 13.1
Q4_K_S 4.58 12.7 GB 14.8 GB Won't fit 4K 17.0
Q4_0 4.55 12.6 GB 14.7 GB Won't fit 5K 17.1
IQ4_XS 4.25 11.9 GB 13.9 GB Fits, but tight 10K 23.6
Q3_K_M 3.91 11.0 GB 13.1 GB Fits, but tight 16K 25.4
IQ3_M 3.70 10.4 GB 12.5 GB Runs comfortably 19K 26.6
IQ3_XXS 3.06 8.8 GB 10.9 GB Runs comfortably 30K 31.3
Q2_K 2.63 7.7 GB 9.8 GB Runs comfortably 32K 35.4
IQ2_XXS 2.06 6.2 GB 8.3 GB Runs comfortably 32K 43.0
IQ1_M 1.75 5.4 GB 7.5 GB Runs comfortably 32K 48.7

Also worth checking