TokenAssemble

Microsoft · open-weights model

Phi-3.5-mini-instruct

Phi-3.5-mini-instruct is a 3.8B-parameter model by Microsoft. Its smallest quant (Q4_K_M) needs at least 6.1GB of GPU-addressable memory at 8K context; it loads cleanly at a quality quant on 50 of the 50 machines we track.

Parameters
3.8B
Layers
32
Max context
128K
Attention
MHA
License
mit

Config from microsoft/Phi-3.5-mini-instruct

Quants and what they need

Measured file sizes; “needs” = weights + KV cache at 8K context + overhead.

QuantFile sizeNeeds ≥
Q8_04.1 GB7.7 GB
Q6_K3.1 GB6.8 GB
Q5_K_M2.8 GB6.5 GB
Q4_K_M2.4 GB6.1 GB

What runs Phi-3.5-mini-instruct

Every machine we track, graded at its best clean-fit quant.

HardwareBest-quant verdict · llama.cpp · 8K context
Intel Arc A770 16 GBFits · A+ · Q8_0interactivebeta
Intel Arc B580Fits · A+ · Q8_0interactivebeta
AMD Radeon PRO W7900Fits · A+ · Q8_0interactivebeta
NVIDIA GeForce RTX 3060 12 GBFits · A+ · Q8_0interactive
NVIDIA GeForce RTX 3070Fits · B · Q5_K_Minteractive
NVIDIA GeForce RTX 3070 TiFits · B · Q5_K_Minteractive
NVIDIA GeForce RTX 3080 10 GBFits · B · Q8_0interactive
NVIDIA GeForce RTX 3080 12 GBFits · A+ · Q8_0interactive
NVIDIA GeForce RTX 3080 TiFits · A+ · Q8_0interactive
NVIDIA GeForce RTX 3090Fits · A+ · Q8_0interactive
NVIDIA GeForce RTX 3090 TiFits · A+ · Q8_0interactive
NVIDIA GeForce RTX 4060Fits · B · Q5_K_Minteractive
NVIDIA GeForce RTX 4060 Ti 16 GBFits · A+ · Q8_0interactive
NVIDIA GeForce RTX 4060 Ti 8 GBFits · B · Q5_K_Minteractive
NVIDIA GeForce RTX 4070Fits · A+ · Q8_0interactive
NVIDIA GeForce RTX 4070 SUPERFits · A+ · Q8_0interactive
NVIDIA GeForce RTX 4070 TiFits · A+ · Q8_0interactive
NVIDIA GeForce RTX 4070 Ti SUPERFits · A+ · Q8_0interactive
NVIDIA GeForce RTX 4080Fits · A+ · Q8_0interactive
NVIDIA GeForce RTX 4080 SUPERFits · A+ · Q8_0interactive
NVIDIA GeForce RTX 4090Fits · A+ · Q8_0interactive
NVIDIA RTX 5000 Ada GenerationFits · A+ · Q8_0interactive
NVIDIA GeForce RTX 5060 Ti 16 GBFits · A+ · Q8_0interactive
NVIDIA GeForce RTX 5070Fits · A+ · Q8_0interactive
NVIDIA GeForce RTX 5070 TiFits · A+ · Q8_0interactive
NVIDIA GeForce RTX 5080Fits · A+ · Q8_0interactive
NVIDIA GeForce RTX 5090Fits · A+ · Q8_0interactive
NVIDIA RTX 6000 Ada GenerationFits · A+ · Q8_0interactive
NVIDIA RTX A6000Fits · A+ · Q8_0interactive
AMD Radeon RX 6800Fits · A+ · Q8_0interactivebeta
AMD Radeon RX 6900 XTFits · A+ · Q8_0interactivebeta
AMD Radeon RX 7800 XTFits · A+ · Q8_0interactivebeta
AMD Radeon RX 7900 XTFits · A+ · Q8_0interactivebeta
AMD Radeon RX 7900 XTXFits · A+ · Q8_0interactivebeta
AMD Radeon RX 9070Fits · A+ · Q8_0interactivebeta
AMD Radeon RX 9070 XTFits · A+ · Q8_0interactivebeta
ASUS Ascent GX10Fits · A+ · Q8_0interactivebeta
Beelink GTR9 Pro Ryzen AI Max+ 395 128GBFits · A+ · Q8_0interactivebeta
NVIDIA DGX SparkFits · A+ · Q8_0interactivebeta
Framework Desktop Ryzen AI Max+ 395 128GBFits · A+ · Q8_0interactivebeta
Framework Desktop Ryzen AI Max+ 395 64GBFits · A+ · Q8_0interactivebeta
GMKtec EVO-X2 Ryzen AI Max+ 395 128GBFits · A+ · Q8_0interactivebeta
HP Z2 Mini G1a Ryzen AI Max+ PRO 395 128GBFits · A+ · Q8_0interactivebeta
Mac mini M4 16GBFits · A+ · Q8_0usable
Mac mini M4 24GBFits · A+ · Q8_0usable
Mac mini M4 Pro 24GBFits · A+ · Q8_0interactive
Mac mini M4 Pro 48GBFits · A+ · Q8_0interactive
Mac Studio M3 Ultra 96GBFits · A+ · Q8_0interactive
Mac Studio M4 Max 36GBFits · A+ · Q8_0interactive
Mac Studio M4 Max 64GBFits · A+ · Q8_0interactive

Verdicts computed from published specs and measured GGUF sizes — how the math works.