AI Models for Nothing Phone (3) — What runs on 16GB

21 great · 14 slow · 13 won't fit
CHIP
Snapdragon 8s Gen 4
MEMORY BANDWIDTH
76.8 GB/s
NPU
RAM OPTIONS
12 / 16 GB
USABLE FOR MODELS
~12 GB
YEAR
2025

Specs checked against manufacturer and public documentation on .

What runs on the Nothing Phone (3)

All 48 models at their recommended quant, on the 16GB configuration. Tap a model for the full report.

ModelParamsQuantNeedsSpeedVerdict
SQwen3 0.6B0.6BQ8_01.3 GB~57.6 tokens/s Runs great
SQwen3 1.7B1.7BQ8_02.6 GB~19.2 tokens/s Runs great
SLlama 3.2 1B1.2BQ4_K_M1.5 GB~43.2 tokens/s Runs great
SLlama 3.2 3B3.2BQ4_K_M2.9 GB~17.3 tokens/s Runs great
SGemma 3 1B1BQ4_K_M1.5 GB~43.2 tokens/s Runs great
SDeepSeek R1 Distill 1.5B1.8BQ4_K_M1.9 GB~31.4 tokens/s Runs great
SSmolLM2 1.7B1.7BQ4_K_M1.9 GB~31.4 tokens/s Runs great
SSmolLM3 3B3.1BQ4_K_M2.8 GB~18.2 tokens/s Runs great
SQwen 3.5 2B2BQ4_K_M2.1 GB~26.6 tokens/s Runs great
STernary Bonsai 8B8BPQ2_03.5 GB~15.7 tokens/s Runs great
STernary Bonsai 4B4BPQ2_02 GB~31.4 tokens/s Runs great
STernary Bonsai 1.7B1.7BPQ2_01.2 GB~69.1 tokens/s Runs great
SMinistral 3 3B3BQ4_K_M3 GB~16.5 tokens/s Runs great
SLFM2.5 8B-A1B8BQ4_K_M6.6 GB~53.2 tokens/s Runs great
SQwen3 4B4BQ4_K_M3.5 GB~13.8 tokens/s Runs great
SGemma 3 4B4.3BQ4_K_M3.5 GB~13.8 tokens/s Runs great
SPhi-4 Mini 3.8B3.8BQ4_K_M3.5 GB~13.8 tokens/s Runs great
SQwen 3.5 4B4BQ4_K_M3.7 GB~12.8 tokens/s Runs great
SNemotron 3 Nano 4B4BQ4_K_M3.8 GB~12.3 tokens/s Runs great
SGemma 4 E2B2BQ4_K_M4 GB~11.1 tokens/s Runs great
ABonsai 27B (1-bit)27BQ1_06.5 GB~9.1 tokens/s Runs great
AMistral 7B v0.37.2BQ4_K_M5.7 GB~7.9 tokens/s! Runs, barely
ADeepSeek R1 Distill 7B7.6BQ4_K_M6.1 GB~7.4 tokens/s! Runs, barely
ALlama 3.1 8B8BQ4_K_M6.3 GB~7.1 tokens/s! Runs, barely
AMinistral 8B8BQ4_K_M6.3 GB~7.1 tokens/s! Runs, barely
AQwen3 8B8.2BQ4_K_M6.4 GB~6.9 tokens/s! Runs, barely
AGemma 4 E4B4BQ4_K_M6.1 GB~6.9 tokens/s! Runs, barely
AMinistral 3 8B8BQ4_K_M6.6 GB~6.6 tokens/s! Runs, barely
AQwen 3.5 9B9BQ4_K_M7.2 GB~6.1 tokens/s! Runs, barely
AGemma 3 12B12.2BQ4_K_M9.1 GB~4.7 tokens/s! Runs, barely
AGemma 4 12B12BQ4_08.8 GB~4.9 tokens/s! Runs, barely
ATernary Bonsai 27B27BPQ2_010.1 GB~4.8 tokens/s! Runs, barely
BMinistral 3 14B14BQ4_K_M10.2 GB~4.2 tokens/s! Runs, barely
BQwen3 14B14.8BQ4_K_M11.1 GB~3.8 tokens/s! Runs, barely
BPhi-4 14B14.7BQ4_K_M11.2 GB~3.8 tokens/s! Runs, barely
FGPT-OSS 20B21BMXFP414.8 GB Won't fit
FGemma 4 26B-A4B26BQ4_017.5 GB Won't fit
FQwen 3.5 27B27BQ4_K_M20 GB Won't fit
FQwen 3.6 27B27BQ4_K_M20.1 GB Won't fit
FGemma 4 31B31BQ4_K_M22 GB Won't fit
FQwen3 30B A3B30.5BQ4_K_M22.3 GB Won't fit
FQwen3 32B32.8BQ4_K_M23.7 GB Won't fit
FQwen 3.5 35B-A3B35BQ4_K_M26.2 GB Won't fit
FQwen 3.6 35B-A3B35BQ4_K_M26.3 GB Won't fit
FNemotron 3 Nano 30B-A3B30BQ4_K_M28.5 GB Won't fit
FLlama 3.3 70B70BQ4_K_M50.1 GB Won't fit
FHunyuan 3 (Hy3)298.8BQ4_K_M212.8 GB Won't fit
FInkling952.4BQ8_0966.9 GB Won't fit

~ = bandwidth-based estimate · ✓ = measured on real hardware

Best model by use case

BEST FOR CHAT

Top everyday assistant & writing pick here — ~57.6 tokens/s at Q8_0, using 1.3 of ~12GB.

BEST FOR CODING

Top code completion & explain-this pick here — ~13.8 tokens/s at Q4_K_M, using 3.5 of ~12GB.

BEST FOR REASONING

Top math & step-by-step thinking pick here — ~31.4 tokens/s at Q4_K_M, using 1.9 of ~12GB.

FAQ

What is the biggest AI model the Nothing Phone (3) can run?

Bonsai 27B (1-bit) (27B parameters) at Q1_0 — it needs 6.5GB of the ~12GB usable on the 16GB Nothing Phone (3) and runs at ~9.1 tokens/s.

How much of the Nothing Phone (3)'s 16GB RAM can AI models actually use?

About 12GB. Android keeps roughly 2–4GB for the system and resident apps, so of the 16GB about 12GB is actually available to a model.

Can the Nothing Phone (3) run Llama 3.1 8B?

Yes — at Q4_K_M it needs 6.3GB of the ~12GB usable and runs at ~7.1 tokens/s.

How fast is local AI on the Nothing Phone (3)?

The Snapdragon 8s Gen 4 has 76.8GB/s of memory bandwidth, which is what decode speed scales with. Small models like Ternary Bonsai 1.7B reach ~69.1 tokens/s; larger 7–14B models land in the single digits. Anything above ~8 tokens/s feels smooth for chat.

Which quantization should I use on the Nothing Phone (3)?

Q4_K_M is the size/quality sweet spot for most models. For example, Qwen3 0.6B at Q8_0 takes 1.3GB of memory here. Only drop to Q3 or IQ4 if a model just misses fitting; Q8 rarely pays off on 16GB of RAM.

Is 16GB of RAM enough for local AI?

35 of the 48 models we track fit on the Nothing Phone (3) — 21 run great and 14 run with compromises. 13 models (mostly 12B+) don't fit at their recommended quant.

Other Nothing phones

Nothing Phone (2a)Nothing Phone (3a)Nothing Phone (4a) ProCMF Phone 2 Pro