AI Compatibility Tier List
NVIDIA A100 80GB
374 AI models ranked by how well they run on your 80 GB VRAM hardware. Best match: Qwen3-Coder-Next (score: 97).
374
Models tested
330
Can run
80 GB
VRAM
116
Best tok/s
Tier Distribution
S 26
A 65
B 53
C 186
F 44
S: 26A: 65B: 53C: 186F: 44
S
S Tier -- Perfect Fit (26 models)
These models run flawlessly with headroom to spare.
Qwen3-Coder-Next80B59.2 GB116 tok/s244K ctx
moe
Qwen 2.5 VL 72B72B57.7 GB42 tok/s33K ctx
dense
Qwen 3.6 35B A3B35B34.4 GB218 tok/s194K ctx
moe
Qwen3-Coder 30B A3B Instruct30.5B29.0 GB259 tok/s256K ctx
moe
Qwen 3.5 27B27B28.5 GB112 tok/s131K ctx
dense
Qwen 3.5 35B A3B35B31.7 GB237 tok/s131K ctx
moe
Qwen3-VL 30B A3B Instruct30B28.7 GB268 tok/s256K ctx
moe
Qwen 3 32B32B32.3 GB95 tok/s131K ctx
dense
Qwen 3.6 27B27B26.3 GB70 tok/s262K ctx
dense
Magistral Small 250724B26.0 GB126 tok/s131K ctx
dense
Devstral Small 2 24B Instruct24B26.0 GB126 tok/s256K ctx
dense
Qwen 3 30B A3B30.5B29.0 GB259 tok/s131K ctx
moe
Nemotron 3 Nano 30B30B29.6 GB101 tok/s131K ctx
dense
Command A 111B111B80.5 GB23 tok/s14K ctx
dense
Llama 3.3 70B70B56.5 GB44 tok/s93K ctx
dense
Gemma 4 31B30.7B42.3 GB60 tok/s57K ctx
dense
Devstral Small 1.124B26.0 GB126 tok/s131K ctx
dense
Qwen 3.5 9B9B16.6 GB126 tok/s131K ctx
dense
Qwen 3 14B14B19.9 GB196 tok/s131K ctx
dense
Nemotron Cascade 2 30B A3B30B30.1 GB265 tok/s262K ctx
moe
Phi-4-reasoning-plus 14B14.7B20.9 GB205 tok/s33K ctx
dense
GPT-OSS 20B21B24.2 GB329 tok/s128K ctx
moe
Qwen 3 8B8B16.0 GB112 tok/s131K ctx
dense
Llama 3.1 70B70B56.5 GB44 tok/s93K ctx
dense
EXAONE 4.0 32B32B32.3 GB95 tok/s131K ctx
dense
Qwen 2.5 72B72B57.7 GB42 tok/s89K ctx
dense
A
A Tier -- Great Fit (65 models)
Excellent performance with comfortable memory margins.
Qwen 3.5 122B A10B122B85.8 GB52 tok/s4K ctx
moe
Gemma 4 26B A4B25.2B27.9 GB278 tok/s243K ctx
moe
Qwen 2.5 32B32B32.3 GB95 tok/s131K ctx
dense
Qwen 3.5 4B4B13.5 GB56 tok/s131K ctx
dense
Kimi Linear 48B A3B48B39.1 GB59 tok/s722K ctx
linear
Gemma 3 27B27B36.6 GB68 tok/s78K ctx
dense
OLMo 2 32B32B32.3 GB95 tok/s4K ctx
dense
Mistral Small 3.2 24B24B26.0 GB126 tok/s131K ctx
vision
Mistral Small 4 119B119B86.9 GB56 tok/s4K ctx
moe
Codestral 2 25.0822B24.8 GB123 tok/s256K ctx
dense
Granite 4.1 30B30B31.1 GB101 tok/s131K ctx
dense
Ministral 3 14B14B19.9 GB196 tok/s262K ctx
multimodal
LFM2 24B24B26.0 GB126 tok/s131K ctx
dense
Devstral 2 123B Instruct123B89.3 GB18 tok/s4K ctx
dense
CogVLM2 19B19B22.9 GB159 tok/s8K ctx
dense
Nemotron Nano 8B8B15.7 GB112 tok/s131K ctx
dense
Mistral Small 24B24B26.0 GB126 tok/s33K ctx
dense
DeepSeek R1 Distill 70B70B56.5 GB44 tok/s93K ctx
dense
Llama 4 Scout 17B 16E109B78.3 GB66 tok/s25K ctx
moe
Phi-4 Mini Reasoning 4B3.8B12.7 GB53 tok/s131K ctx
dense
Mistral Small 3.1 24B24B26.0 GB126 tok/s131K ctx
dense
GPT-OSS 120B117B85.2 GB20 tok/s4K ctx
dense
Phi-4 14B14B20.5 GB196 tok/s16K ctx
dense
InternVL2 8B8B15.7 GB112 tok/s8K ctx
dense
Qwen 2.5 Coder 32B32B32.3 GB95 tok/s131K ctx
dense
Qwen 2.5 14B14B20.4 GB196 tok/s131K ctx
dense
Pixtral Large 124B124B89.9 GB17 tok/s4K ctx
dense
Granite Code 20B20B24.3 GB152 tok/s8K ctx
dense
Leanstral 119B A6B119B90.3 GB48 tok/s4K ctx
moe
Granite Code 34B34B33.3 GB90 tok/s8K ctx
dense
Gemma 3 12B12B21.1 GB153 tok/s131K ctx
dense
Command R 35B35B32.7 GB87 tok/s131K ctx
dense
Ministral 3 8B8B16.0 GB112 tok/s262K ctx
multimodal
MiniCPM-V 2.6 8B8B15.7 GB112 tok/s2K ctx
dense
MPT-30B-Instruct30B53.9 GB81 tok/s8K ctx
dense
DeepSeek Coder V2 16B16B22.0 GB418 tok/s131K ctx
moe
StarCoder 15B15B34.3 GB162 tok/s8K ctx
dense
Nemotron 70B70B56.5 GB44 tok/s93K ctx
dense
DeepSeek R1 Distill 32B32B32.3 GB95 tok/s33K ctx
dense
Nemotron Nano 9B v29B16.8 GB126 tok/s131K ctx
dense
SQLCoder 7B7B15.1 GB98 tok/s8K ctx
dense
Qwen 3 4B4B13.5 GB56 tok/s33K ctx
dense
Qwen 2.5 VL 7B7B14.0 GB98 tok/s33K ctx
dense
CodeLlama 13B Instruct13B29.0 GB182 tok/s16K ctx
dense
Jina Embeddings v30.57B12.0 GB8 tok/s8K ctx
dense
OLMo 2 13B13B19.3 GB182 tok/s33K ctx
dense
Magistral 7B7B15.1 GB98 tok/s8K ctx
dense
CodeGeeX 4 9B9B15.0 GB126 tok/s131K ctx
dense
Gemma 4 E4B8B15.1 GB112 tok/s128K ctx
dense
BGE M30.57B11.2 GB8 tok/s8K ctx
dense
LLaVA 1.6 13B13B29.0 GB182 tok/s4K ctx
dense
Falcon 40B Instruct40B39.5 GB66 tok/s8K ctx
dense
DeepSeek R1 Distill 14B14B20.4 GB196 tok/s33K ctx
dense
Qwen 2.5 7B7B14.0 GB98 tok/s131K ctx
dense
Granite Code 8B8B15.7 GB112 tok/s8K ctx
dense
DevStral 7B7B15.1 GB98 tok/s8K ctx
dense
CodeLlama 7B Instruct7B21.0 GB98 tok/s16K ctx
dense
StarCoder 7B7B20.5 GB98 tok/s8K ctx
dense
Nous Dolphin 13B13B30.5 GB182 tok/s16K ctx
dense
Codestral Mamba 7B7B13.7 GB98 tok/s262K ctx
state-space
Pixtral 12B12B18.7 GB168 tok/s131K ctx
dense
Granite 4.1 8B8B16.2 GB112 tok/s131K ctx
dense
Gemma 2 27B27B36.6 GB68 tok/s8K ctx
dense
WizardLM 13B13B29.0 GB182 tok/s8K ctx
dense
Vicuna 13B13B29.0 GB182 tok/s4K ctx
dense
B
B Tier -- Good Fit (53 models)
Solid performance, may need quantization or reduced context.
mxbai Embed Large0.34B11.1 GB5 tok/s1K ctx
dense
Nous Hermes 1.09B26.6 GB126 tok/s16K ctx
dense
Snowflake Arctic Embed L0.34B11.1 GB5 tok/s1K ctx
dense
Qwen 2.5 Coder 3B3B12.9 GB42 tok/s131K ctx
dense
InternLM 7B7B21.0 GB98 tok/s8K ctx
dense
InternLM Chat 7B7B21.0 GB98 tok/s8K ctx
dense
Mixtral 8x7B47B39.5 GB123 tok/s33K ctx
moe
Qwen 2.5 Math 72B72B57.7 GB42 tok/s4K ctx
dense
Command R+ 104B104B75.8 GB29 tok/s36K ctx
dense
Gemma 4 E2B5.1B12.5 GB71 tok/s128K ctx
dense
GLM-4 9B9B15.0 GB126 tok/s128K ctx
dense
Llama 3.1 8B8B15.7 GB112 tok/s128K ctx
dense
BGE Large EN v1.50.34B11.1 GB5 tok/s1K ctx
dense
Solar 7B7B16.1 GB98 tok/s8K ctx
dense
WizardMath 7B7B15.1 GB98 tok/s4K ctx
dense
OLMo 2 7B7B15.1 GB98 tok/s4K ctx
dense
Baichuan 13B13B30.5 GB182 tok/s8K ctx
dense
Qwen 2.5 Coder 7B7B14.0 GB98 tok/s131K ctx
dense
Cerebras-GPT 13B13B28.0 GB182 tok/s101K ctx
dense
Nomic Embed Text v1.50.14B9.7 GB2 tok/s8K ctx
dense
LLaVA 1.5 7B7B21.0 GB98 tok/s4K ctx
dense
Gemma 3 4B4B13.4 GB56 tok/s128K ctx
dense
Ministral 3 3B3B11.5 GB42 tok/s262K ctx
multimodal
DeepSeek LLM 67B67B55.6 GB46 tok/s4K ctx
dense
Phi 4 Mini 4B4B12.8 GB56 tok/s128K ctx
dense
MPT-7B-Instruct7B21.0 GB98 tok/s8K ctx
dense
Baichuan 7B7B21.0 GB98 tok/s8K ctx
dense
Qwen 3.5 2B2B11.8 GB28 tok/s131K ctx
dense
Yi 1.5 34B34B33.3 GB90 tok/s4K ctx
dense
DeepSeek R1 Distill 8B8B15.7 GB112 tok/s33K ctx
dense
DeepSeek R1 Distill 7B7B14.0 GB98 tok/s33K ctx
dense
Samantha 7B7B15.1 GB98 tok/s4K ctx
dense
Phi 3 Mini 3.8B3.8B17.1 GB53 tok/s128K ctx
dense
Qwen 2.5 3B3B12.9 GB42 tok/s131K ctx
dense
Qwen 2.5 Coder 14B14B20.4 GB196 tok/s131K ctx
dense
Falcon 7B Instruct7B13.3 GB98 tok/s8K ctx
dense
Gemma 2 9B9B19.5 GB126 tok/s8K ctx
dense
Llama 3.2 11B Vision11B17.6 GB154 tok/s16K ctx
vision
InternLM 20B20B41.6 GB121 tok/s8K ctx
dense
Phi 3.5 Mini 4B4B17.2 GB56 tok/s128K ctx
dense
Qwen 3 1.7B1.7B11.6 GB24 tok/s33K ctx
dense
All MiniLM L6 v20.02B9.2 GB2 tok/s0K ctx
dense
Granite 4.1 3B3B12.0 GB42 tok/s131K ctx
dense
Mistral Nemo 12B12B18.7 GB168 tok/s128K ctx
dense
Granite Code 3B3B13.2 GB42 tok/s8K ctx
dense
Qwen 2.5 Coder 1.5B1.5B10.2 GB21 tok/s33K ctx
dense
Codestral 22B22B24.8 GB130 tok/s33K ctx
dense
Phi 3 Medium 14B14B20.5 GB196 tok/s128K ctx
dense
Yi Coder 9B9B15.9 GB126 tok/s131K ctx
dense
Mistral 7B Instruct v0.37B15.1 GB98 tok/s8K ctx
dense
Ministral 8B8B16.0 GB112 tok/s131K ctx
dense
Aya Expanse 32B32B30.9 GB95 tok/s8K ctx
dense
Llama 3.2 3B3B12.4 GB42 tok/s128K ctx
dense
C
C Tier -- Tight Fit (186 models)
Runs with offloading or significant compromises.
Llama 3.3 70B Instruct
C5470B59.8 GB40 tok/s55K ctx
dense
Solar Open 69B REAP i1
C5469B59.1 GB41 tok/s57K ctx
dense
stabilityai japanese stablelm instruct beta 70b
C5470B59.8 GB40 tok/s55K ctx
dense
BaichuanMed OCR 72B i1
C5472B61.3 GB39 tok/s52K ctx
dense
Yi 34B Chat34B33.3 GB90 tok/s200K ctx
dense
Qwen3 48B A4B Savant Commander Distill 12X Closed Open Heretic Uncensored
C5248B43.8 GB59 tok/s119K ctx
dense
Qwen3.5 35B A3B35B34.4 GB80 tok/s194K ctx
dense
Qwen3.5 35B A3B
C5135B34.4 GB80 tok/s194K ctx
dense
Granite 3.1 8B8B15.7 GB112 tok/s128K ctx
state-space
Hermes 4.3 36B
C5136B35.1 GB78 tok/s186K ctx
dense
DeepSeek R1 1.5B1.5B10.2 GB21 tok/s33K ctx
dense
Yi 1.5 9B9B15.9 GB126 tok/s4K ctx
dense
SmolLM3 3B3B12.7 GB42 tok/s128K ctx
dense
TinyLlama 1.1B1.1B9.9 GB15 tok/s4K ctx
dense
EXAONE 4.0 32B
C5032B32.2 GB88 tok/s220K ctx
dense
aya expanse 32b heretic MPOA i1
C5032B32.2 GB88 tok/s220K ctx
dense
Baichuan M2 32B Q4 K M
C5032B32.2 GB88 tok/s220K ctx
dense
baichuan inc Baichuan M2 32B
C5032B32.2 GB88 tok/s220K ctx
dense
Qwen3.5 27B27B28.5 GB104 tok/s276K ctx
dense
Qwen 2.5 Math 7B7B14.0 GB98 tok/s4K ctx
dense
gemma 3 27b it27B28.5 GB104 tok/s276K ctx
dense
StableLM 2 12B12B29.7 GB152 tok/s4K ctx
dense
gemma 3 27b it
C5027B28.5 GB104 tok/s276K ctx
dense
Qwen3.5 122B A10B122B83.0 GB21 tok/s13K ctx
dense
Solar Open 100B
C49100B81.6 GB23 tok/s14K ctx
dense
Solar Open 100B i1
C49100B81.6 GB23 tok/s14K ctx
dense
Mistral Small 3.2 24B Instruct 250624B26.4 GB117 tok/s321K ctx
dense
cognitivecomputations Dolphin Mistral 24B Venice Edition
C4924B26.4 GB117 tok/s321K ctx
dense
Mistral Small 24B Instruct 2501
C4924B26.4 GB117 tok/s321K ctx
dense
mistral small 3.1 24b instruct 2503 hf
C4924B26.4 GB117 tok/s321K ctx
dense
cognitivecomputations Dolphin3.0 R1 Mistral 24B
C4924B26.4 GB117 tok/s321K ctx
dense
Dolphin Mistral GLM 4.7 Flash 24B Venice Edition Thinking Uncensored i1
C4924B26.4 GB117 tok/s321K ctx
dense
StarCoder2 15B15B20.9 GB177 tok/s16K ctx
dense
cognitivecomputations Dolphin Mistral 24B Venice Edition
C4924B26.4 GB117 tok/s321K ctx
dense
Codestral 22B v0.1
C4822B24.9 GB128 tok/s358K ctx
dense
Codestral 22B v0.1
C4822B24.9 GB128 tok/s358K ctx
dense
Codestral 22B v0.1
C4822B24.9 GB128 tok/s358K ctx
dense
Codestral 22B v0.1 IMat
C4822B24.9 GB128 tok/s358K ctx
dense
Vicuna 7B7B21.0 GB98 tok/s4K ctx
dense
Qwen 2.5 1.5B1.5B10.2 GB21 tok/s131K ctx
dense
Codestral 22B v0.1 i1
C4822B24.9 GB128 tok/s358K ctx
dense
OpenChat 7B7B15.1 GB98 tok/s8K ctx
dense
Codestral 21B Pruned i1
C4821B24.2 GB134 tok/s379K ctx
dense
internlm2 5 20b chat
C4820B23.4 GB140 tok/s402K ctx
dense
internlm2 limarp chat 20b
C4820B23.4 GB140 tok/s402K ctx
dense
internlm2 math plus 20b i1
C4820B23.4 GB140 tok/s402K ctx
dense
Aya Expanse 8B8B15.7 GB112 tok/s8K ctx
dense
Codestral RAG 19B Pruned i1
C4719B22.7 GB148 tok/s428K ctx
dense
DeepSeek R1 Distill Qwen 14B14B19.1 GB196 tok/s610K ctx
dense
Phi 4 reasoning vision 15B
C4715B19.8 GB187 tok/s564K ctx
dense
StarCoder2 15B
C4715B19.8 GB187 tok/s564K ctx
dense
NousResearch Hermes 4 14B
C4714B19.1 GB196 tok/s610K ctx
dense
starcoder2 15b instruct v0.1
C4715B19.8 GB187 tok/s564K ctx
dense
starcoder2 15b i1
C4715B19.8 GB187 tok/s564K ctx
dense
starcoder2 15b instruct v0.1
C4715B19.8 GB187 tok/s564K ctx
dense
Dolphin 2.9 8B8B15.7 GB112 tok/s33K ctx
dense
HelpingAI 15B i1
C4715B19.8 GB187 tok/s564K ctx
dense
OpenHermes 2.5 7B7B15.1 GB98 tok/s8K ctx
dense
gemma 3 12b it
C4712B17.6 GB168 tok/s726K ctx
dense
GGUF SOLARized GraniStral 14B 2102 YeAM HCT 32QKV
C4714B19.1 GB196 tok/s610K ctx
dense
internlm JanusCoder 14B
C4714B19.1 GB196 tok/s610K ctx
dense
GGUF SOLARized GraniStral 14B 1902 YeAM HCT
C4714B19.1 GB196 tok/s610K ctx
dense
Gemma 3 1B1B9.9 GB14 tok/s33K ctx
dense
Zephyr 7B Beta7B15.1 GB98 tok/s33K ctx
dense
Qwen3.5 9B9B15.4 GB126 tok/s995K ctx
dense
Qwen3.5 9B Uncensored HauhauCS Aggressive
C469B15.4 GB126 tok/s995K ctx
dense
SOLAR 10.7B Instruct v1.0 uncensored
C4610.7B16.7 GB150 tok/s824K ctx
dense
Nous Hermes 2 SOLAR 10.7B
C4610.7B16.7 GB150 tok/s824K ctx
dense
Qwen3.5 9B
C469B15.4 GB126 tok/s995K ctx
dense
Starling LM 7B7B15.1 GB98 tok/s8K ctx
dense
Meta Llama 3.1 8B Instruct
C468B14.7 GB112 tok/s1130K ctx
dense
llava llama 3 8b v1 1
C468B14.7 GB112 tok/s1130K ctx
dense
DeepSeek R1 0528 Qwen3 8B8B14.7 GB112 tok/s1130K ctx
dense
Yi Coder 9B Chat
C469B15.4 GB126 tok/s995K ctx
dense
glm 4 9b chat 1m
C469B15.4 GB126 tok/s995K ctx
dense
Meta Llama 3 8B Instruct
C468B14.7 GB112 tok/s1130K ctx
dense
DeepSeek R1 Distill Llama 8B8B14.7 GB112 tok/s1130K ctx
dense
solar finalised finetuned Model 10.7B i1
C4610.7B16.7 GB150 tok/s824K ctx
dense
Dolphin3.0 Llama3.1 8B
C468B14.7 GB112 tok/s1130K ctx
dense
Llama 3 8B Instruct 32k v0.1
C468B14.7 GB112 tok/s1130K ctx
dense
Meta Llama 3.1 8B Instruct
C468B14.7 GB112 tok/s1130K ctx
dense
vntl llama3 8b v2
C468B14.7 GB112 tok/s1130K ctx
dense
SOLAR 10.7B v1.0
C4610.7B16.7 GB150 tok/s824K ctx
dense
DeepSeek R1 0528 Qwen3 8B
C468B14.7 GB112 tok/s1130K ctx
dense
Llama 2 7B Chat
C467B14.0 GB98 tok/s1304K ctx
dense
Mistral 7B Instruct v0.2
C467B14.0 GB98 tok/s1304K ctx
dense
Yi 1.5 9B Chat
C469B15.4 GB126 tok/s995K ctx
dense
DeepSeek R1 0528 Qwen3 8B
C468B14.7 GB112 tok/s1130K ctx
dense
Gemma 2 2B2B11.7 GB28 tok/s8K ctx
dense
Mistral 7B Instruct v0.3
C467B14.0 GB98 tok/s1304K ctx
dense
Qwen3 8B DeepSeek v3.2 Speciale Distill
C468B14.7 GB112 tok/s1130K ctx
dense
Hermes 3 Llama 3.1 8B
C468B14.7 GB112 tok/s1130K ctx
dense
Hermes 2 Pro Llama 3 8B
C468B14.7 GB112 tok/s1130K ctx
dense
Dolphin3.0 Llama3.1 8B
C468B14.7 GB112 tok/s1130K ctx
dense
DeepSeek LLM 7B7B20.5 GB98 tok/s4K ctx
dense
granite 8b code instruct 4k
C468B14.7 GB112 tok/s1130K ctx
dense
dolphin 2.9.4 llama3.1 8b
C468B14.7 GB112 tok/s1130K ctx
dense
HelpingAI2.5 10B i1
C4610B16.2 GB140 tok/s887K ctx
dense
Helply 10.2b chat i1
C4610.2B16.3 GB143 tok/s868K ctx
dense
Yi 9B Coder i1
C469B15.4 GB126 tok/s995K ctx
dense
openchat 3.6 8b 20240522 IMat
C468B14.7 GB112 tok/s1130K ctx
dense
Mistral 7B Instruct v0.3
C467B14.0 GB98 tok/s1304K ctx
dense
Hermes 2 Pro Mistral 7B
C467B14.0 GB98 tok/s1304K ctx
dense
aya expanse 8b
C468B14.7 GB112 tok/s1130K ctx
dense
EXAONE 3.5 7.8B Instruct
C467.8B14.6 GB109 tok/s1161K ctx
dense
HelpingAI2 9B
C469B15.4 GB126 tok/s995K ctx
dense
zephyr 7B beta
C467B14.0 GB98 tok/s1304K ctx
dense
dolphin v2 8b abliterated i1
C468B14.7 GB112 tok/s1130K ctx
dense
Nous Hermes 2 Mistral 7B DPO
C467B14.0 GB98 tok/s1304K ctx
dense
EXAONE 3.5 7.8B Instruct
C467.8B14.6 GB109 tok/s1161K ctx
dense
EXAONE 3.5 7.8B Instruct i1
C467.8B14.6 GB109 tok/s1161K ctx
dense
HelpingAI2 9B i1
C469B15.4 GB126 tok/s995K ctx
dense
aya expanse 8b orthogonal heretic i1
C468B14.7 GB112 tok/s1130K ctx
dense
Falcon H1R 7B
C467B14.0 GB98 tok/s1304K ctx
dense
HelpingAI 9B 200k i1
C459B15.4 GB126 tok/s995K ctx
dense
exaone 3.0 7.8b it
C457.8B14.6 GB109 tok/s1161K ctx
dense
falcon mamba 7b instruct Q4 K M
C457B14.0 GB98 tok/s1304K ctx
dense
HelpingAI 9B i1
C459B15.4 GB126 tok/s995K ctx
dense
zephyr 7B alpha
C457B14.0 GB98 tok/s1304K ctx
dense
speechless zephyr code functionary 7b
C457B14.0 GB98 tok/s1304K ctx
dense
StarCoder2 7B
C457B14.0 GB98 tok/s1304K ctx
dense
aya expanse 8b orthogonal heretic
C458B14.7 GB112 tok/s1130K ctx
dense
japanese stablelm instruct gamma 7B
C457B14.0 GB98 tok/s1304K ctx
dense
Falcon H1 7B Instruct
C457B14.0 GB98 tok/s1304K ctx
dense
Falcon H1R 7B7B14.0 GB98 tok/s1304K ctx
dense
internlm2 5 1 8b chat i1
C458B14.7 GB112 tok/s1130K ctx
dense
internlm3 8b instruct abliterated i1
C458B14.7 GB112 tok/s1130K ctx
dense
Mamba Codestral 7B v0.1
C457B14.0 GB98 tok/s1304K ctx
dense
Yi 1.5 6B Chat
C456B13.3 GB84 tok/s1535K ctx
dense
stabilityai japanese stablelm base gamma 7b
C457B14.0 GB98 tok/s1304K ctx
dense
baichuan2 7b chat
C457B14.0 GB98 tok/s1304K ctx
dense
Qwen 2.5 Coder 0.5B0.5B9.4 GB7 tok/s131K ctx
dense
DiscoPOP zephyr 7b gemma
C457B14.0 GB98 tok/s1304K ctx
dense
starcoder2 7b
C457B14.0 GB98 tok/s1304K ctx
dense
Neural Chat 7B7B15.1 GB98 tok/s8K ctx
dense
stablelm 2 zephyr 1 6b
C456B13.3 GB84 tok/s1535K ctx
dense
zephyr 7b beta Mistral 7B Instruct v0.2
C457B14.0 GB98 tok/s1304K ctx
dense
OpenChat 3.5 7B Qwen v2.0 i1
C457B14.0 GB98 tok/s1304K ctx
dense
OpenChat 3.5 7B Starling v2.0 i1
C457B14.0 GB98 tok/s1304K ctx
dense
internlm2 math plus 7b IMat
C457B14.0 GB98 tok/s1304K ctx
dense
CodeNinja 1.0 OpenChat 7B i1
C457B14.0 GB98 tok/s1304K ctx
dense
internlm2 5 7b chat i1
C457B14.0 GB98 tok/s1304K ctx
dense
OpenSafetyLab MD Judge v0 2 internlm2 7b
C457B14.0 GB98 tok/s1304K ctx
dense
MD Judge v0 2 internlm2 7b i1
C457B14.0 GB98 tok/s1304K ctx
dense
zephyr 7b gemma sft african ultrachat 100k
C457B14.0 GB98 tok/s1304K ctx
dense
jointpreferences mistral 7b sft helpful
C457B14.0 GB98 tok/s1304K ctx
dense
zephyr 7b dpo full i1
C457B14.0 GB98 tok/s1304K ctx
dense
blossom v3 baichuan2 7b i1
C457B14.0 GB98 tok/s1304K ctx
dense
blossom v1 baichuan 7b i1
C457B14.0 GB98 tok/s1304K ctx
dense
Yi 1.5 6B6B13.5 GB84 tok/s4K ctx
dense
Yi 1.5 6B Chat
C456B13.3 GB84 tok/s1535K ctx
dense
stablelm 2 1 6b chat imatrix
C456B13.3 GB84 tok/s1535K ctx
dense
StarCoder2 7B7B13.7 GB98 tok/s16K ctx
dense
HelpingAI2 6B
C456B13.3 GB84 tok/s1535K ctx
dense
HelpingAI2 6B i1
C456B13.3 GB84 tok/s1535K ctx
dense
Nemotron Mini 4B4B13.3 GB56 tok/s4K ctx
dense
HelpingAI2.5 5B i1
C445B12.5 GB70 tok/s1858K ctx
dense
Qwen3.5 4B4B11.8 GB56 tok/s2344K ctx
dense
gemma 3 4b it
C444B11.8 GB56 tok/s2344K ctx
dense
gemma 3 4b it
C444B11.8 GB56 tok/s2344K ctx
dense
Qwen 3 0.6B0.6B10.1 GB8 tok/s33K ctx
dense
Llama 3.2 1B1B10.0 GB14 tok/s128K ctx
dense
Llama 3.2 3B Instruct
C433B11.4 GB42 tok/s3138K ctx
dense
Qwen2.5 3B Instruct3B11.1 GB42 tok/s3153K ctx
dense
Llama 3.2 3B Instruct
C423B11.1 GB42 tok/s3153K ctx
dense
Ministral 3 3B Instruct 2512
C423B11.1 GB42 tok/s3153K ctx
dense
HELVETE 3B
C423B11.1 GB42 tok/s3153K ctx
dense
Hermes 3 Llama 3.2 3B
C423B11.1 GB42 tok/s3153K ctx
dense
stablelm zephyr 3b
C423B11.1 GB42 tok/s3153K ctx
dense
StarCoder2 3B
C423B11.1 GB42 tok/s3153K ctx
dense
AI21 Jamba Reasoning 3B
C423B11.1 GB42 tok/s3153K ctx
dense
stablelm 3b 4e1t
C423B11.1 GB42 tok/s3153K ctx
dense
ai21labs AI21 Jamba Reasoning 3B
C423B11.1 GB42 tok/s3153K ctx
dense
ai21labs AI21 Jamba2 3B
C423B11.1 GB42 tok/s3153K ctx
dense
HelpingAI 3B hindi i1
C423B11.1 GB42 tok/s3153K ctx
dense
AI21 Jamba2 3B
C423B11.1 GB42 tok/s3153K ctx
dense
HelpingAI 3B hindi
C423B11.1 GB42 tok/s3153K ctx
dense
AI21 Jamba2 3B i1
C423B11.1 GB42 tok/s3153K ctx
dense
gemma 2 2b it
C412B10.8 GB28 tok/s4742K ctx
dense
EXAONE 3.5 2.4B Instruct
C412.4B10.6 GB34 tok/s3962K ctx
dense
gemma 2b2B10.4 GB28 tok/s4770K ctx
dense
Gemmasutra Mini 2B v1
C412B10.4 GB28 tok/s4770K ctx
dense
gemma 2 2b it
C412B10.4 GB28 tok/s4770K ctx
dense
Qwen 3.5 0.6B0.6B10.1 GB8 tok/s131K ctx
dense
StarCoder2 3B3B11.2 GB42 tok/s16K ctx
dense
Qwen2.5 1.5B Instruct1.5B10.0 GB21 tok/s6388K ctx
dense
F
F Tier -- Can't Run (44 models)
Too heavy for this hardware.
DeepSeek R1 Distill Qwen 1.5B1.5B10.0 GB21 tok/s6388K ctx
dense
Yi Coder 1.5B Chat
D401.5B10.0 GB21 tok/s6388K ctx
dense
stablelm 2 zephyr 1.6b
D401.6B10.1 GB22 tok/s5984K ctx
dense
logos16v2 stablelm2 1.6b i1
D401.6B10.1 GB22 tok/s5984K ctx
dense
Yi Coder 1.5B
D401.5B10.0 GB21 tok/s6388K ctx
dense
Falcon H1 1.5B Instruct1.5B10.0 GB21 tok/s6388K ctx
dense
TinyLlama 1.1B Chat v1.0
D391.1B9.7 GB15 tok/s8742K ctx
dense
Qwen 2.5 0.5B0.5B9.4 GB7 tok/s131K ctx
dense
EXAONE 4.0 1.2B
D391.2B9.8 GB17 tok/s8006K ctx
dense
Llama 3.2 1B Instruct Q8 0
D391B9.8 GB14 tok/s9596K ctx
dense
TinyLlama 1.1B Chat v0.3
D391.1B9.7 GB15 tok/s8742K ctx
dense
TinyLlama 1.1B Chat v0.61.1B9.7 GB15 tok/s8742K ctx
dense
Llama 3.2 1B Instruct
D391B9.6 GB14 tok/s9624K ctx
dense
gemma 3 1b it
D391B9.6 GB14 tok/s9624K ctx
dense
TinyLlama 1.1B Chat v1.0 imatrix
D381.1B9.7 GB15 tok/s8742K ctx
dense
Falcon3 1B Instruct abliterated
D381B9.6 GB14 tok/s9624K ctx
dense
SmolVLM 500M Instruct
D370.5B9.4 GB7 tok/s11310K ctx
dense
embeddinggemma 300M
D350.3B9.2 GB4 tok/s11337K ctx
dense
granite embedding 107m multilingual
D330.11B9.1 GB2 tok/s11366K ctx
dense
Falcon H1 Tiny 90M Instruct
D330.09B9.1 GB2 tok/s11367K ctx
dense
Qwen 3 235B A22B235B155.1 GB10 tok/s4K ctx
moe
Qwen3-Coder 480B A35B Instruct480B304.6 GB3 tok/s4K ctx
moe
Qwen 3.5 397B A17B397B253.9 GB4 tok/s4K ctx
moe
Mixtral 8x22B141B98.3 GB25 tok/s4K ctx
moe
Mistral Large 3675B426.0 GB2 tok/s4K ctx
moe
GLM-5744B481.8 GB2 tok/s4K ctx
moe
GLM-5.1754B487.9 GB2 tok/s4K ctx
moe
Kimi K2.51000B626.3 GB2 tok/s4K ctx
moe
Kimi K2.61000B626.3 GB2 tok/s4K ctx
moe
DeepSeek V3 671B671B477.8 GB2 tok/s4K ctx
moe
DeepSeek V3.1 671B671B477.8 GB2 tok/s4K ctx
moe
DeepSeek V4 Pro1600B872.8 GB2 tok/s4K ctx
moe
DeepSeek V4 Flash284B168.2 GB9 tok/s4K ctx
moe
DeepSeek R1 671B671B477.8 GB2 tok/s4K ctx
moe
DeepSeek Coder V2 236B236B211.5 GB6 tok/s4K ctx
moe
DeepSeek V2.5 236B236B211.5 GB6 tok/s4K ctx
moe
Llama 4 Maverick 17B 128E400B255.8 GB3 tok/s4K ctx
moe
Llama 3.1 405B405B263.6 GB2 tok/s4K ctx
dense
DeepSeek V3.2671B418.7 GB2 tok/s4K ctx
moe
MiniMax M2.7
F0230B153.0 GB12 tok/s4K ctx
moe
K EXAONE 236B A23B
F0236B180.5 GB3 tok/s4K ctx
dense
Baichuan M3 235B
F0235B179.8 GB3 tok/s4K ctx
dense
Baichuan M3 235B i1
F0235B179.8 GB3 tok/s4K ctx
dense
Qwen3.5 397B A17B397B297.6 GB2 tok/s4K ctx
dense
Compare with...
See How Other Hardware Stacks Up
Check tier lists for similar GPUs and Apple Silicon configs