mistralai · text

Mistral-Small-Instruct-2409

mistralai/Mistral-Small-Instruct-2409

Mistral-Small-Instruct-2409 at Q4_K_M is exactly 13,341,242,112 bytes (12.42 GiB / 13.34 GB) — an effective 4.797 bits per weight, not the nominal 4. Its KV cache at 32K is 7.00 GiB.

From the file· summed from 1 file(s)From the file· KV per layer
Parameters
22.2B
Architecture
llama
56 layers
Context
32,768
native (config.json)
License
—

Shipped quantizations

● exact bytes, summed from published files
QuantSize●Exact bytes●Effective bpw●Tensors●Publisher
IQ1_S4.50 GiB4,829,489,2801.737—MaziyarPanahi
IQ1_M4.91 GiB5,267,138,6881.894—MaziyarPanahi
IQ2_XXS5.58 GiB5,996,557,3122.156—bartowski
IQ2_XS6.19 GiB6,646,147,2002.390—MaziyarPanahi
IQ2_XS6.19 GiB6,646,150,1442.390—bartowski
IQ2_M7.10 GiB7,618,966,5282.740—bartowski
Q2_K7.70 GiB8,272,098,0482.975—MaziyarPanahi
Q2_K7.70 GiB8,272,098,3042.975—bartowski
Q2_K_L7.89 GiB8,468,706,3043.045—bartowski
IQ3_XS8.55 GiB9,176,098,9443.300—MaziyarPanahi
IQ3_XS8.55 GiB9,176,101,8883.300—bartowski
Q3_K_S8.98 GiB9,641,276,1603.467—MaziyarPanahi
Q3_K_S8.98 GiB9,641,276,4163.467—bartowski
IQ3_M9.37 GiB10,062,410,7523.618—bartowski
Q3_K_M10.02 GiB10,756,829,9523.868—MaziyarPanahi
Q3_K_M10.02 GiB10,756,830,2083.868507bartowski
Q3_K_L10.92 GiB11,730,432,7684.218—MaziyarPanahi
Q3_K_L10.92 GiB11,730,433,0244.218—bartowski
IQ4_XS11.12 GiB11,935,295,6164.292—MaziyarPanahi
IQ4_XS11.12 GiB11,935,298,5604.292507bartowski
Q4_011.75 GiB12,613,202,9444.536507bartowski
Q4_K_S11.79 GiB12,660,388,6084.553—MaziyarPanahi
Q4_K_S11.79 GiB12,660,388,8644.553—bartowski
Q4_K_M12.42 GiB13,341,242,1124.797—MaziyarPanahi
Q4_K_M12.43 GiB13,341,242,3684.797507bartowski
Q4_K_L12.56 GiB13,490,664,4484.851—bartowski
Q5_K_S14.27 GiB15,324,820,2245.511—MaziyarPanahi
Q5_K_S14.27 GiB15,324,820,4805.511—bartowski
Q5_K_M14.64 GiB15,722,558,2085.654—MaziyarPanahi
Q5_K_M14.64 GiB15,722,558,4645.654507bartowski
Q5_K_L14.76 GiB15,846,814,7205.698—bartowski
Q6_K17.00 GiB18,252,703,8726.564—MaziyarPanahi
Q6_K17.00 GiB18,252,706,5606.564—MaziyarPanahi
Q6_K17.00 GiB18,252,706,8166.564507bartowski
Q6_K_L17.09 GiB18,350,224,3846.599—bartowski
IQ1_S2 shards18.79 GiB20,172,977,1207.254—MaziyarPanahi
IQ1_M2 shards20.51 GiB22,018,339,8087.918—MaziyarPanahi
Q8_022.02 GiB23,640,549,5048.501—MaziyarPanahi
Q8_022.02 GiB23,640,552,1928.501—MaziyarPanahi
Q8_022.02 GiB23,640,552,4488.501507bartowski

KV cache by context

computed per layer
ContextKV cache (f16)●Flat formulaOverstated byFull / windowed / recurrent
4,0960.88 GiB0.88 GiB—56 / 0 / 0
8,1921.75 GiB1.75 GiB—56 / 0 / 0
16,3843.50 GiB3.50 GiB—56 / 0 / 0
32,7687.00 GiB7.00 GiB—56 / 0 / 0
65,53614.00 GiB14.00 GiB—56 / 0 / 0
131,07228.00 GiB28.00 GiB—56 / 0 / 0

Compare with

same modality, comparable size
vs Qwen3-Coder-30B-A3B-Instructvs Qwen3.6-27Bvs Qwen3.8-27Bvs gemma-4-12B-it-qat-q4_0-unquantizedvs gemma-4-E4B-itvs Qwen3-30B-A3B-Thinking-2507All alternatives →Will it fit on my card? →

Will it run on your card?

full quant x context sweep
By VRAM:8GB12GB16GB24GB32GB48GB
Radeon RX 6500 XT 4GBGeForce RTX 3050 6GBGeForce RTX 5050 8GBGeForce RTX 3080 10GBGeForce RTX 2080 Ti 11GBGeForce RTX 5070 12GBGeForce RTX 5060 Ti 16GBApple M3 Pro 18GBGeForce RTX 3080 Ti 20GBGeForce RTX 5090 D V2 24GBGeForce RTX 5090 D 32GBApple M5 Max 36GBApple M5 Pro 48GBApple M5 Max 64GBApple M3 Ultra 96GBApple M5 Max 128GBApple M2 Ultra 192GBApple M3 Ultra 256GBApple M3 Ultra 512GB

Why other calculators give a different number

A parameters × bits ÷ 8 estimate puts Q4_K_M at roughly 11.65 GiB. The real file is 12.42 GiB, because a quantization is a mixture and some tensors are always kept at higher precision.

Architecture

from config.json
Layers
56
Attention heads
48
KV heads
8
Head dim
128
Hidden size
6144
Vocab
32,768
Sliding window
none
SWA period
—
MLA
no
Experts
—
Experts per token
—
use_sliding_window
—

Questions people ask

How much VRAM does Mistral-Small-Instruct-2409 need?
Q4_K_M is exactly 13,341,242,112 bytes (12.42 GiB) in weights. Add the KV cache, which depends on your context length, plus roughly half a gigabyte of runtime overhead.
How large is Mistral-Small-Instruct-2409's KV cache?
7.00 GiB at 32K context with an f16 cache, computed per layer. Quantizing the cache to q8_0 roughly halves it, which is often the difference between a context length fitting and not.
Which quantization of Mistral-Small-Instruct-2409 should I use?
Q4_K_M is the usual default. Pick the largest quantization that fits your card at the context you actually need — the table above gives exact sizes for every one published.