llamaperf

Best local LLMs by hardware tier

Rankings only make sense once you fix the hardware. Pick a tier below — the leaderboard re-sorts to the models the community actually runs there, weighted by report count, fastest observed tokens-per-second, and recency.

Looking for a coding model? See the best local LLM for coding by VRAM.

Hardware tier
Model size

NVIDIA 16 GB

Midrange NVIDIA cards. Handles 13B comfortably and 30B-class with tight quants. Ranked from 65 reports.

#Model familyReportsFastest t/s
1Qwen3.8Alibaba
27B · 27B · NVFP4 · on V100 16GB
48219.1
2Qwen3.6Alibaba
35B-A3B · 35B-A3B · IQ4_XS-4.19bpw · on RTX 4070 Ti Super
11110.2
3DeepSeek V4 FlashDeepSeek
284B-A13B · 284B-A13B · on RTX 5060 Ti 16GB
211.0
4Agents-A1InternScience
Uncensored-MTP-APEX · 35B-A3B · APEX Compact · on RTX 5070 Ti
197.0
5Gemma 4Google DeepMind
E4B · 7.5B · Q4_K_M · on RTX 5060 Ti 16GB
173.9
6MuseMeta
Glimmer · 30B · Q4_K_XL · on RTX 5060 Ti 16GB
118.0
7Qwen3Alibaba
30B-A3B · 30B-A3B · float8 · on RTX 5060 Ti 16GB
152.0

How we rank

A single global "best models" list doesn't really exist — what runs well on a 5090 is often unrunnable on a 4060, and a 7B that screams on an M3 Max is usually a poor pick on an H100. So we fix the hardware first, then rank the families that actually have community reports on it. The score blends popularity (log-scaled report count), fastest observed tokens-per-second normalized within the bucket, recency (90-day half-life), and a small bias for rows where we know the variant + quant + GPU cleanly. Click into a family for the full breakdown of records.