HuggingFaceTB · text

SmolLM-1.7B-Instruct-v0.2

HuggingFaceTB/SmolLM-1.7B-Instruct-v0.2

SmolLM-1.7B-Instruct-v0.2 at Q4_K_M is exactly 1,055,610,144 bytes (0.98 GiB / 1.06 GB) — an effective 4.935 bits per weight, not the nominal 4.

From the file· summed from 1 file(s)
Parameters
1.7B
Architecture
llama
Context
—
native (config.json)
License
apache-2.0

Shipped quantizations

● exact bytes, summed from published files
QuantSize●Exact bytes●Effective bpw●Tensors●Publisher
IQ1_S0.38 GiB412,308,7681.927—MaziyarPanahi
IQ1_M0.41 GiB444,159,2642.076—MaziyarPanahi
IQ2_XS0.51 GiB542,856,4802.538—MaziyarPanahi
Q2_K0.63 GiB674,583,8403.153—MaziyarPanahi
Q2_K0.63 GiB674,583,9043.153—bartowski
IQ3_XXS0.63 GiB680,088,9283.179—bartowski
Q2_K_L0.65 GiB698,963,2963.267—bartowski
IQ3_XS0.69 GiB739,071,2643.455—MaziyarPanahi
IQ3_XS0.69 GiB739,071,3283.455—bartowski
Q3_K_S0.72 GiB776,820,0003.631—MaziyarPanahi
IQ3_M0.75 GiB810,243,4243.788—bartowski
Q3_K_M0.80 GiB860,181,7924.021—MaziyarPanahi
Q3_K_L0.87 GiB932,533,5364.359—MaziyarPanahi
Q3_K_L0.87 GiB932,533,6004.359—bartowski
IQ4_XS0.88 GiB940,397,8564.396—MaziyarPanahi
IQ4_XS0.88 GiB940,397,9204.396—bartowski
Q4_K_S0.93 GiB999,118,1124.670—MaziyarPanahi
Q4_K_S0.93 GiB999,118,1764.670—bartowski
Q4_K_M0.98 GiB1,055,610,1444.935—MaziyarPanahi
Q4_K_M0.98 GiB1,055,610,2084.935—bartowski
Q4_K_L1.01 GiB1,079,989,6005.048—bartowski
Q5_K_S1.11 GiB1,192,056,0965.572—MaziyarPanahi
Q5_K_S1.11 GiB1,192,056,1605.572—bartowski
Q5_K_M1.14 GiB1,225,479,4565.729—MaziyarPanahi
Q5_K_M1.14 GiB1,225,479,5205.729—bartowski
Q5_K_L1.16 GiB1,249,858,9125.843—bartowski
Q6_K1.31 GiB1,405,965,6006.572—MaziyarPanahi
Q6_K1.31 GiB1,405,965,6646.572—bartowski
Q6_K_L1.33 GiB1,430,345,0566.686—bartowski
Q8_01.70 GiB1,820,415,2648.510—MaziyarPanahi
Q8_01.70 GiB1,820,415,3288.510—bartowski
F326.38 GiB6,847,288,38432.008—bartowski

Compare with

same modality, comparable size
vs Llama-3.2-1B-Instructvs ced-basevs gemma-3-1b-itvs Qwen3-1.7Bvs embeddinggemma-300mvs Llama-3.2-3B-InstructAll 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 0.90 GiB. The real file is 0.98 GiB, because a quantization is a mixture and some tensors are always kept at higher precision.

Architecture

Architecture unavailable — this repository is gated and no ungated mirror was found. Exact file sizes above are still authoritative; only the KV math needs the config.

Questions people ask

How much VRAM does SmolLM-1.7B-Instruct-v0.2 need?
Q4_K_M is exactly 1,055,610,144 bytes (0.98 GiB) in weights. Add the KV cache, which depends on your context length, plus roughly half a gigabyte of runtime overhead.
Which quantization of SmolLM-1.7B-Instruct-v0.2 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.