AI Compatibility Tier List
NVIDIA H200 PCIe 141GB
374 AI models ranked by how well they run on your 141 GB VRAM hardware. Best match: Qwen 3.5 122B A10B (score: 98).
374
Models tested
332
Can run
141 GB
VRAM
162
Best tok/s
Tier Distribution
S 29
A 57
B 58
C 188
F 42
S: 29A: 57B: 58C: 188F: 42
S
S Tier -- Perfect Fit (29 models)
These models run flawlessly with headroom to spare.
Qwen 3.5 122B A10B122B91.9 GB162 tok/s131K ctx
moe
Devstral 2 123B Instruct123B95.4 GB58 tok/s152K ctx
dense
Mistral Small 4 119B119B93.0 GB176 tok/s159K ctx
moe
GPT-OSS 120B117B91.3 GB61 tok/s131K ctx
dense
Pixtral Large 124B124B96.0 GB58 tok/s131K ctx
dense
Command A 111B111B86.6 GB65 tok/s239K ctx
dense
Leanstral 119B A6B119B96.4 GB162 tok/s97K ctx
moe
Qwen 2.5 VL 72B72B63.8 GB100 tok/s33K ctx
dense
Qwen3-Coder-Next80B65.3 GB272 tok/s256K ctx
moe
Qwen3-Coder 30B A3B Instruct30.5B35.1 GB610 tok/s256K ctx
moe
Qwen 3.6 35B A3B35B40.5 GB512 tok/s262K ctx
moe
Qwen 3.5 27B27B34.6 GB264 tok/s131K ctx
dense
Qwen3-VL 30B A3B Instruct30B34.8 GB631 tok/s256K ctx
moe
Qwen 3.6 27B27B32.4 GB165 tok/s262K ctx
dense
Qwen 3.5 35B A3B35B37.8 GB557 tok/s131K ctx
moe
Qwen 3 32B32B38.4 GB225 tok/s131K ctx
dense
Magistral Small 250724B32.1 GB296 tok/s131K ctx
dense
Devstral Small 2 24B Instruct24B32.1 GB296 tok/s256K ctx
dense
Qwen 3 30B A3B30.5B35.1 GB610 tok/s131K ctx
moe
Nemotron 3 Nano 30B30B35.7 GB237 tok/s131K ctx
dense
Qwen 3.5 9B9B22.7 GB126 tok/s131K ctx
dense
Qwen 3 14B14B26.0 GB196 tok/s131K ctx
dense
Devstral Small 1.124B32.1 GB296 tok/s131K ctx
dense
Gemma 4 31B30.7B48.4 GB141 tok/s117K ctx
dense
Llama 3.3 70B70B62.6 GB103 tok/s128K ctx
dense
Phi-4-reasoning-plus 14B14.7B27.0 GB206 tok/s33K ctx
dense
Qwen 3 8B8B22.1 GB112 tok/s131K ctx
dense
Nemotron Cascade 2 30B A3B30B36.2 GB623 tok/s262K ctx
moe
GPT-OSS 20B21B30.3 GB774 tok/s128K ctx
moe
A
A Tier -- Great Fit (57 models)
Excellent performance with comfortable memory margins.
Qwen 3.5 4B4B19.6 GB56 tok/s131K ctx
dense
EXAONE 4.0 32B32B38.4 GB223 tok/s131K ctx
dense
Qwen 2.5 72B72B63.8 GB100 tok/s131K ctx
dense
Llama 4 Scout 17B 16E109B84.4 GB154 tok/s325K ctx
moe
Llama 3.1 70B70B62.6 GB103 tok/s128K ctx
dense
Gemma 4 26B A4B25.2B34.0 GB655 tok/s256K ctx
moe
Qwen 2.5 32B32B38.4 GB223 tok/s131K ctx
dense
Kimi Linear 48B A3B48B45.2 GB138 tok/s1049K ctx
linear
Gemma 3 27B27B42.7 GB160 tok/s131K ctx
dense
Ministral 3 14B14B26.0 GB196 tok/s262K ctx
multimodal
Codestral 2 25.0822B30.9 GB288 tok/s256K ctx
dense
Mistral Small 3.2 24B24B32.1 GB296 tok/s131K ctx
vision
Nemotron Nano 8B8B21.8 GB112 tok/s131K ctx
dense
Qwen 3 235B A22B235B161.2 GB48 tok/s4K ctx
moe
Phi-4 Mini Reasoning 4B3.8B18.8 GB53 tok/s131K ctx
dense
OLMo 2 32B32B38.4 GB223 tok/s4K ctx
dense
LFM2 24B24B32.1 GB296 tok/s131K ctx
dense
Granite 4.1 30B30B37.2 GB237 tok/s131K ctx
dense
CogVLM2 19B19B29.0 GB266 tok/s8K ctx
dense
Mistral Small 24B24B32.1 GB296 tok/s33K ctx
dense
MiniMax M2.7
A79230B159.1 GB56 tok/s4K ctx
moe
Mistral Small 3.1 24B24B32.1 GB296 tok/s131K ctx
dense
DeepSeek R1 Distill 70B70B62.6 GB103 tok/s131K ctx
dense
Phi-4 14B14B26.6 GB196 tok/s16K ctx
dense
InternVL2 8B8B21.8 GB112 tok/s8K ctx
dense
Qwen 2.5 14B14B26.5 GB196 tok/s131K ctx
dense
Granite Code 20B20B30.4 GB280 tok/s8K ctx
dense
Ministral 3 8B8B22.1 GB112 tok/s262K ctx
multimodal
MiniCPM-V 2.6 8B8B21.8 GB112 tok/s2K ctx
dense
Gemma 3 12B12B27.2 GB168 tok/s131K ctx
dense
Qwen 2.5 Coder 32B32B38.4 GB223 tok/s131K ctx
dense
Nemotron Nano 9B v29B22.9 GB126 tok/s131K ctx
dense
Granite Code 34B34B39.4 GB211 tok/s8K ctx
dense
DeepSeek Coder V2 16B16B28.1 GB984 tok/s131K ctx
moe
SQLCoder 7B7B21.2 GB98 tok/s8K ctx
dense
Qwen 3 4B4B19.6 GB56 tok/s33K ctx
dense
Qwen 2.5 VL 7B7B20.1 GB98 tok/s33K ctx
dense
Command R 35B35B38.8 GB205 tok/s131K ctx
dense
Jina Embeddings v30.57B18.1 GB8 tok/s8K ctx
dense
StarCoder 15B15B40.4 GB210 tok/s8K ctx
dense
DeepSeek R1 Distill 32B32B38.4 GB223 tok/s33K ctx
dense
Nemotron 70B70B62.6 GB103 tok/s131K ctx
dense
Magistral 7B7B21.2 GB98 tok/s8K ctx
dense
CodeGeeX 4 9B9B21.1 GB126 tok/s131K ctx
dense
BGE M30.57B17.3 GB8 tok/s8K ctx
dense
OLMo 2 13B13B25.4 GB182 tok/s33K ctx
dense
Gemma 4 E4B8B21.2 GB112 tok/s128K ctx
dense
CodeLlama 13B Instruct13B35.1 GB182 tok/s16K ctx
dense
MPT-30B-Instruct30B60.0 GB190 tok/s8K ctx
dense
Qwen 2.5 7B7B20.1 GB98 tok/s131K ctx
dense
Command R+ 104B104B81.9 GB69 tok/s131K ctx
dense
LLaVA 1.6 13B13B35.1 GB182 tok/s4K ctx
dense
DeepSeek R1 Distill 14B14B26.5 GB196 tok/s33K ctx
dense
Mixtral 8x22B141B104.4 GB97 tok/s66K ctx
moe
Granite Code 8B8B21.8 GB112 tok/s8K ctx
dense
DevStral 7B7B21.2 GB98 tok/s8K ctx
dense
Codestral Mamba 7B7B19.8 GB98 tok/s262K ctx
state-space
B
B Tier -- Good Fit (58 models)
Solid performance, may need quantization or reduced context.
Granite 4.1 8B8B22.3 GB112 tok/s131K ctx
dense
Falcon 40B Instruct40B45.6 GB155 tok/s8K ctx
dense
mxbai Embed Large0.34B17.2 GB5 tok/s1K ctx
dense
CodeLlama 7B Instruct7B27.1 GB98 tok/s16K ctx
dense
StarCoder 7B7B26.6 GB98 tok/s8K ctx
dense
Pixtral 12B12B24.8 GB168 tok/s131K ctx
dense
Snowflake Arctic Embed L0.34B17.2 GB5 tok/s1K ctx
dense
Nous Dolphin 13B13B36.6 GB182 tok/s16K ctx
dense
Gemma 2 27B27B42.7 GB160 tok/s8K ctx
dense
Qwen 2.5 Coder 3B3B19.0 GB42 tok/s131K ctx
dense
WizardLM 13B13B35.1 GB182 tok/s8K ctx
dense
Vicuna 13B13B35.1 GB182 tok/s4K ctx
dense
Nous Hermes 1.09B32.7 GB126 tok/s16K ctx
dense
Gemma 4 E2B5.1B18.6 GB71 tok/s128K ctx
dense
BGE Large EN v1.50.34B17.2 GB5 tok/s1K ctx
dense
InternLM 7B7B27.1 GB98 tok/s8K ctx
dense
InternLM Chat 7B7B27.1 GB98 tok/s8K ctx
dense
GLM-4 9B9B21.1 GB126 tok/s128K ctx
dense
Llama 3.1 8B8B21.8 GB112 tok/s128K ctx
dense
Solar 7B7B22.2 GB98 tok/s8K ctx
dense
WizardMath 7B7B21.2 GB98 tok/s4K ctx
dense
OLMo 2 7B7B21.2 GB98 tok/s4K ctx
dense
Nomic Embed Text v1.50.14B15.8 GB2 tok/s8K ctx
dense
Qwen 2.5 Math 72B72B63.8 GB100 tok/s4K ctx
dense
Qwen 2.5 Coder 7B7B20.1 GB98 tok/s131K ctx
dense
Gemma 3 4B4B19.5 GB56 tok/s128K ctx
dense
Ministral 3 3B3B17.6 GB42 tok/s262K ctx
multimodal
Mixtral 8x7B47B45.6 GB290 tok/s33K ctx
moe
LLaVA 1.5 7B7B27.1 GB98 tok/s4K ctx
dense
Phi 4 Mini 4B4B18.9 GB56 tok/s128K ctx
dense
Baichuan 13B13B36.6 GB182 tok/s8K ctx
dense
Cerebras-GPT 13B13B34.1 GB182 tok/s131K ctx
dense
Qwen 3.5 2B2B17.9 GB28 tok/s131K ctx
dense
MPT-7B-Instruct7B27.1 GB98 tok/s8K ctx
dense
Baichuan 7B7B27.1 GB98 tok/s8K ctx
dense
DeepSeek R1 Distill 7B7B20.1 GB98 tok/s33K ctx
dense
DeepSeek R1 Distill 8B8B21.8 GB112 tok/s33K ctx
dense
Samantha 7B7B21.2 GB98 tok/s4K ctx
dense
DeepSeek LLM 67B67B61.7 GB107 tok/s4K ctx
dense
Qwen 2.5 3B3B19.0 GB42 tok/s131K ctx
dense
Falcon 7B Instruct7B19.4 GB98 tok/s8K ctx
dense
Phi 3 Mini 3.8B3.8B23.2 GB53 tok/s128K ctx
dense
Yi 1.5 34B34B39.4 GB211 tok/s4K ctx
dense
Gemma 2 9B9B25.6 GB126 tok/s8K ctx
dense
Qwen 2.5 Coder 14B14B26.5 GB196 tok/s131K ctx
dense
Qwen 3 1.7B1.7B17.7 GB24 tok/s33K ctx
dense
Llama 3.2 11B Vision11B23.7 GB154 tok/s16K ctx
vision
All MiniLM L6 v20.02B15.3 GB2 tok/s0K ctx
dense
Phi 3.5 Mini 4B4B23.3 GB56 tok/s128K ctx
dense
Granite 4.1 3B3B18.1 GB42 tok/s131K ctx
dense
Qwen 2.5 Coder 1.5B1.5B16.3 GB21 tok/s33K ctx
dense
Granite Code 3B3B19.3 GB42 tok/s8K ctx
dense
Mistral Nemo 12B12B24.8 GB168 tok/s128K ctx
dense
Yi Coder 9B9B22.0 GB126 tok/s131K ctx
dense
Mistral 7B Instruct v0.37B21.2 GB98 tok/s8K ctx
dense
InternLM 20B20B47.7 GB280 tok/s8K ctx
dense
Phi 3 Medium 14B14B26.6 GB196 tok/s128K ctx
dense
Codestral 22B22B30.9 GB307 tok/s33K ctx
dense
C
C Tier -- Tight Fit (188 models)
Runs with offloading or significant compromises.
Ministral 8B8B22.1 GB112 tok/s131K ctx
dense
Llama 3.2 3B3B18.5 GB42 tok/s128K ctx
dense
Qwen3.5 122B A10B122B89.1 GB63 tok/s74K ctx
dense
Solar Open 100B
C54100B87.7 GB66 tok/s89K ctx
dense
Solar Open 100B i1
C54100B87.7 GB66 tok/s89K ctx
dense
Aya Expanse 32B32B37.0 GB225 tok/s8K ctx
dense
Llama 3.3 70B Instruct
C5270B65.9 GB94 tok/s162K ctx
dense
BaichuanMed OCR 72B i1
C5272B67.4 GB92 tok/s156K ctx
dense
Solar Open 69B REAP i1
C5269B65.2 GB96 tok/s166K ctx
dense
stabilityai japanese stablelm instruct beta 70b
C5170B65.9 GB94 tok/s162K ctx
dense
DeepSeek R1 1.5B1.5B16.3 GB21 tok/s33K ctx
dense
TinyLlama 1.1B1.1B16.0 GB15 tok/s4K ctx
dense
SmolLM3 3B3B18.8 GB42 tok/s128K ctx
dense
Granite 3.1 8B8B21.8 GB112 tok/s128K ctx
state-space
Yi 1.5 9B9B22.0 GB126 tok/s4K ctx
dense
Yi 34B Chat34B39.4 GB211 tok/s200K ctx
dense
Qwen3 48B A4B Savant Commander Distill 12X Closed Open Heretic Uncensored
C4948B49.9 GB138 tok/s275K ctx
dense
Qwen 2.5 Math 7B7B20.1 GB98 tok/s4K ctx
dense
Qwen3.5 35B A3B35B40.5 GB189 tok/s408K ctx
dense
Qwen3.5 35B A3B
C4835B40.5 GB189 tok/s408K ctx
dense
Hermes 4.3 36B
C4836B41.2 GB184 tok/s395K ctx
dense
Qwen 2.5 1.5B1.5B16.3 GB21 tok/s131K ctx
dense
Qwen3.5 27B27B34.6 GB245 tok/s554K ctx
dense
EXAONE 4.0 32B
C4732B38.3 GB207 tok/s454K ctx
dense
aya expanse 32b heretic MPOA i1
C4732B38.3 GB207 tok/s454K ctx
dense
gemma 3 27b it27B34.6 GB245 tok/s554K ctx
dense
Baichuan M2 32B Q4 K M
C4732B38.3 GB207 tok/s454K ctx
dense
StarCoder2 15B15B27.0 GB210 tok/s16K ctx
dense
baichuan inc Baichuan M2 32B
C4732B38.3 GB207 tok/s454K ctx
dense
gemma 3 27b it
C4727B34.6 GB245 tok/s554K ctx
dense
OpenChat 7B7B21.2 GB98 tok/s8K ctx
dense
StableLM 2 12B12B35.8 GB168 tok/s4K ctx
dense
Mistral Small 3.2 24B Instruct 250624B32.5 GB275 tok/s634K ctx
dense
cognitivecomputations Dolphin Mistral 24B Venice Edition
C4724B32.5 GB275 tok/s634K ctx
dense
Mistral Small 24B Instruct 2501
C4724B32.5 GB275 tok/s634K ctx
dense
mistral small 3.1 24b instruct 2503 hf
C4724B32.5 GB275 tok/s634K ctx
dense
cognitivecomputations Dolphin3.0 R1 Mistral 24B
C4724B32.5 GB275 tok/s634K ctx
dense
Aya Expanse 8B8B21.8 GB112 tok/s8K ctx
dense
Dolphin Mistral GLM 4.7 Flash 24B Venice Edition Thinking Uncensored i1
C4724B32.5 GB275 tok/s634K ctx
dense
Vicuna 7B7B27.1 GB98 tok/s4K ctx
dense
Codestral 22B v0.1
C4722B31.0 GB300 tok/s699K ctx
dense
Codestral 22B v0.1
C4722B31.0 GB300 tok/s699K ctx
dense
cognitivecomputations Dolphin Mistral 24B Venice Edition
C4624B32.5 GB275 tok/s634K ctx
dense
Codestral 22B v0.1
C4622B31.0 GB300 tok/s699K ctx
dense
Gemma 3 1B1B16.0 GB14 tok/s33K ctx
dense
Codestral 22B v0.1 IMat
C4622B31.0 GB300 tok/s699K ctx
dense
Codestral 22B v0.1 i1
C4622B31.0 GB300 tok/s699K ctx
dense
Codestral 21B Pruned i1
C4621B30.3 GB294 tok/s736K ctx
dense
internlm2 5 20b chat
C4620B29.5 GB280 tok/s777K ctx
dense
internlm2 limarp chat 20b
C4620B29.5 GB280 tok/s777K ctx
dense
OpenHermes 2.5 7B7B21.2 GB98 tok/s8K ctx
dense
internlm2 math plus 20b i1
C4620B29.5 GB280 tok/s777K ctx
dense
Dolphin 2.9 8B8B21.8 GB112 tok/s33K ctx
dense
DeepSeek R1 Distill Qwen 14B14B25.2 GB196 tok/s1146K ctx
dense
Phi 4 reasoning vision 15B
C4615B25.9 GB210 tok/s1064K ctx
dense
Codestral RAG 19B Pruned i1
C4619B28.8 GB266 tok/s822K ctx
dense
StarCoder2 15B
C4615B25.9 GB210 tok/s1064K ctx
dense
Zephyr 7B Beta7B21.2 GB98 tok/s33K ctx
dense
NousResearch Hermes 4 14B
C4614B25.2 GB196 tok/s1146K ctx
dense
gemma 3 12b it
C4612B23.7 GB168 tok/s1350K ctx
dense
Qwen3.5 9B9B21.5 GB126 tok/s1828K ctx
dense
Qwen3.5 9B Uncensored HauhauCS Aggressive
C459B21.5 GB126 tok/s1828K ctx
dense
starcoder2 15b instruct v0.1
C4515B25.9 GB210 tok/s1064K ctx
dense
starcoder2 15b i1
C4515B25.9 GB210 tok/s1064K ctx
dense
Gemma 2 2B2B17.8 GB28 tok/s8K ctx
dense
starcoder2 15b instruct v0.1
C4515B25.9 GB210 tok/s1064K ctx
dense
Starling LM 7B7B21.2 GB98 tok/s8K ctx
dense
Meta Llama 3.1 8B Instruct
C458B20.8 GB112 tok/s2067K ctx
dense
Qwen3.5 9B
C459B21.5 GB126 tok/s1828K ctx
dense
GGUF SOLARized GraniStral 14B 2102 YeAM HCT 32QKV
C4514B25.2 GB196 tok/s1146K ctx
dense
HelpingAI 15B i1
C4515B25.9 GB210 tok/s1064K ctx
dense
llava llama 3 8b v1 1
C458B20.8 GB112 tok/s2067K ctx
dense
internlm JanusCoder 14B
C4514B25.2 GB196 tok/s1146K ctx
dense
DeepSeek R1 0528 Qwen3 8B8B20.8 GB112 tok/s2067K ctx
dense
GGUF SOLARized GraniStral 14B 1902 YeAM HCT
C4514B25.2 GB196 tok/s1146K ctx
dense
SOLAR 10.7B Instruct v1.0 uncensored
C4510.7B22.8 GB150 tok/s1524K ctx
dense
Nous Hermes 2 SOLAR 10.7B
C4510.7B22.8 GB150 tok/s1524K ctx
dense
Meta Llama 3 8B Instruct
C458B20.8 GB112 tok/s2067K ctx
dense
Llama 2 7B Chat
C457B20.1 GB98 tok/s2374K ctx
dense
DeepSeek R1 Distill Llama 8B8B20.8 GB112 tok/s2067K ctx
dense
Dolphin3.0 Llama3.1 8B
C458B20.8 GB112 tok/s2067K ctx
dense
Llama 3 8B Instruct 32k v0.1
C458B20.8 GB112 tok/s2067K ctx
dense
Yi Coder 9B Chat
C459B21.5 GB126 tok/s1828K ctx
dense
Mistral 7B Instruct v0.2
C457B20.1 GB98 tok/s2374K ctx
dense
Meta Llama 3.1 8B Instruct
C458B20.8 GB112 tok/s2067K ctx
dense
vntl llama3 8b v2
C458B20.8 GB112 tok/s2067K ctx
dense
glm 4 9b chat 1m
C459B21.5 GB126 tok/s1828K ctx
dense
DeepSeek R1 0528 Qwen3 8B
C458B20.8 GB112 tok/s2067K ctx
dense
Mistral 7B Instruct v0.3
C457B20.1 GB98 tok/s2374K ctx
dense
DeepSeek R1 0528 Qwen3 8B
C458B20.8 GB112 tok/s2067K ctx
dense
Qwen3 8B DeepSeek v3.2 Speciale Distill
C458B20.8 GB112 tok/s2067K ctx
dense
Qwen 2.5 Coder 0.5B0.5B15.5 GB7 tok/s131K ctx
dense
Hermes 3 Llama 3.1 8B
C458B20.8 GB112 tok/s2067K ctx
dense
Yi 1.5 9B Chat
C459B21.5 GB126 tok/s1828K ctx
dense
solar finalised finetuned Model 10.7B i1
C4510.7B22.8 GB150 tok/s1524K ctx
dense
Hermes 2 Pro Llama 3 8B
C458B20.8 GB112 tok/s2067K ctx
dense
Dolphin3.0 Llama3.1 8B
C458B20.8 GB112 tok/s2067K ctx
dense
granite 8b code instruct 4k
C458B20.8 GB112 tok/s2067K ctx
dense
SOLAR 10.7B v1.0
C4510.7B22.8 GB150 tok/s1524K ctx
dense
dolphin 2.9.4 llama3.1 8b
C458B20.8 GB112 tok/s2067K ctx
dense
Mistral 7B Instruct v0.3
C457B20.1 GB98 tok/s2374K ctx
dense
Hermes 2 Pro Mistral 7B
C457B20.1 GB98 tok/s2374K ctx
dense
zephyr 7B beta
C457B20.1 GB98 tok/s2374K ctx
dense
Nous Hermes 2 Mistral 7B DPO
C457B20.1 GB98 tok/s2374K ctx
dense
openchat 3.6 8b 20240522 IMat
C458B20.8 GB112 tok/s2067K ctx
dense
aya expanse 8b
C458B20.8 GB112 tok/s2067K ctx
dense
EXAONE 3.5 7.8B Instruct
C457.8B20.7 GB109 tok/s2122K ctx
dense
Falcon H1R 7B
C457B20.1 GB98 tok/s2374K ctx
dense
Yi 9B Coder i1
C459B21.5 GB126 tok/s1828K ctx
dense
dolphin v2 8b abliterated i1
C458B20.8 GB112 tok/s2067K ctx
dense
EXAONE 3.5 7.8B Instruct
C457.8B20.7 GB109 tok/s2122K ctx
dense
falcon mamba 7b instruct Q4 K M
C457B20.1 GB98 tok/s2374K ctx
dense
zephyr 7B alpha
C457B20.1 GB98 tok/s2374K ctx
dense
speechless zephyr code functionary 7b
C457B20.1 GB98 tok/s2374K ctx
dense
EXAONE 3.5 7.8B Instruct i1
C457.8B20.7 GB109 tok/s2122K ctx
dense
HelpingAI2.5 10B i1
C4510B22.3 GB140 tok/s1637K ctx
dense
aya expanse 8b orthogonal heretic i1
C458B20.8 GB112 tok/s2067K ctx
dense
Helply 10.2b chat i1
C4510.2B22.4 GB143 tok/s1603K ctx
dense
StarCoder2 7B
C457B20.1 GB98 tok/s2374K ctx
dense
exaone 3.0 7.8b it
C457.8B20.7 GB109 tok/s2122K ctx
dense
HelpingAI2 9B
C459B21.5 GB126 tok/s1828K ctx
dense
japanese stablelm instruct gamma 7B
C457B20.1 GB98 tok/s2374K ctx
dense
Falcon H1 7B Instruct
C457B20.1 GB98 tok/s2374K ctx
dense
Falcon H1R 7B7B20.1 GB98 tok/s2374K ctx
dense
HelpingAI2 9B i1
C459B21.5 GB126 tok/s1828K ctx
dense
HelpingAI 9B 200k i1
C459B21.5 GB126 tok/s1828K ctx
dense
aya expanse 8b orthogonal heretic
C458B20.8 GB112 tok/s2067K ctx
dense
Yi 1.5 6B Chat
C456B19.4 GB84 tok/s2784K ctx
dense
HelpingAI 9B i1
C459B21.5 GB126 tok/s1828K ctx
dense
Mamba Codestral 7B v0.1
C457B20.1 GB98 tok/s2374K ctx
dense
stablelm 2 zephyr 1 6b
C456B19.4 GB84 tok/s2784K ctx
dense
stabilityai japanese stablelm base gamma 7b
C457B20.1 GB98 tok/s2374K ctx
dense
internlm2 5 1 8b chat i1
C458B20.8 GB112 tok/s2067K ctx
dense
internlm3 8b instruct abliterated i1
C458B20.8 GB112 tok/s2067K ctx
dense
baichuan2 7b chat
C457B20.1 GB98 tok/s2374K ctx
dense
DiscoPOP zephyr 7b gemma
C457B20.1 GB98 tok/s2374K ctx
dense
starcoder2 7b
C447B20.1 GB98 tok/s2374K ctx
dense
zephyr 7b beta Mistral 7B Instruct v0.2
C447B20.1 GB98 tok/s2374K ctx
dense
OpenChat 3.5 7B Qwen v2.0 i1
C447B20.1 GB98 tok/s2374K ctx
dense
OpenChat 3.5 7B Starling v2.0 i1
C447B20.1 GB98 tok/s2374K ctx
dense
internlm2 math plus 7b IMat
C447B20.1 GB98 tok/s2374K ctx
dense
CodeNinja 1.0 OpenChat 7B i1
C447B20.1 GB98 tok/s2374K ctx
dense
internlm2 5 7b chat i1
C447B20.1 GB98 tok/s2374K ctx
dense
OpenSafetyLab MD Judge v0 2 internlm2 7b
C447B20.1 GB98 tok/s2374K ctx
dense
MD Judge v0 2 internlm2 7b i1
C447B20.1 GB98 tok/s2374K ctx
dense
zephyr 7b gemma sft african ultrachat 100k
C447B20.1 GB98 tok/s2374K ctx
dense
jointpreferences mistral 7b sft helpful
C447B20.1 GB98 tok/s2374K ctx
dense
Yi 1.5 6B6B19.6 GB84 tok/s4K ctx
dense
Yi 1.5 6B Chat
C446B19.4 GB84 tok/s2784K ctx
dense
zephyr 7b dpo full i1
C447B20.1 GB98 tok/s2374K ctx
dense
blossom v3 baichuan2 7b i1
C447B20.1 GB98 tok/s2374K ctx
dense
stablelm 2 1 6b chat imatrix
C446B19.4 GB84 tok/s2784K ctx
dense
blossom v1 baichuan 7b i1
C447B20.1 GB98 tok/s2374K ctx
dense
Neural Chat 7B7B21.2 GB98 tok/s8K ctx
dense
DeepSeek LLM 7B7B26.6 GB98 tok/s4K ctx
dense
HelpingAI2 6B
C446B19.4 GB84 tok/s2784K ctx
dense
StarCoder2 7B7B19.8 GB98 tok/s16K ctx
dense
HelpingAI2 6B i1
C446B19.4 GB84 tok/s2784K ctx
dense
Nemotron Mini 4B4B19.4 GB56 tok/s4K ctx
dense
HelpingAI2.5 5B i1
C435B18.6 GB70 tok/s3357K ctx
dense
Qwen3.5 4B4B17.9 GB56 tok/s4218K ctx
dense
gemma 3 4b it
C434B17.9 GB56 tok/s4218K ctx
dense
gemma 3 4b it
C434B17.9 GB56 tok/s4218K ctx
dense
Qwen 3 0.6B0.6B16.2 GB8 tok/s33K ctx
dense
Llama 3.2 1B1B16.1 GB14 tok/s128K ctx
dense
Llama 3.2 3B Instruct
C423B17.5 GB42 tok/s5636K ctx
dense
Qwen2.5 3B Instruct3B17.2 GB42 tok/s5651K ctx
dense
Llama 3.2 3B Instruct
C423B17.2 GB42 tok/s5651K ctx
dense
Ministral 3 3B Instruct 2512
C423B17.2 GB42 tok/s5651K ctx
dense
HELVETE 3B
C423B17.2 GB42 tok/s5651K ctx
dense
Hermes 3 Llama 3.2 3B
C423B17.2 GB42 tok/s5651K ctx
dense
stablelm zephyr 3b
C423B17.2 GB42 tok/s5651K ctx
dense
StarCoder2 3B
C423B17.2 GB42 tok/s5651K ctx
dense
AI21 Jamba Reasoning 3B
C423B17.2 GB42 tok/s5651K ctx
dense
stablelm 3b 4e1t
C423B17.2 GB42 tok/s5651K ctx
dense
ai21labs AI21 Jamba Reasoning 3B
C423B17.2 GB42 tok/s5651K ctx
dense
ai21labs AI21 Jamba2 3B
C423B17.2 GB42 tok/s5651K ctx
dense
HelpingAI 3B hindi i1
C413B17.2 GB42 tok/s5651K ctx
dense
AI21 Jamba2 3B
C413B17.2 GB42 tok/s5651K ctx
dense
HelpingAI 3B hindi
C413B17.2 GB42 tok/s5651K ctx
dense
AI21 Jamba2 3B i1
C413B17.2 GB42 tok/s5651K ctx
dense
EXAONE 3.5 2.4B Instruct
C412.4B16.7 GB34 tok/s7085K ctx
dense
gemma 2 2b it
C412B16.9 GB28 tok/s8490K ctx
dense
gemma 2b2B16.5 GB28 tok/s8518K ctx
dense
Gemmasutra Mini 2B v1
C412B16.5 GB28 tok/s8518K ctx
dense
Qwen 3.5 0.6B0.6B16.2 GB8 tok/s131K ctx
dense
gemma 2 2b it
C412B16.5 GB28 tok/s8518K ctx
dense
StarCoder2 3B3B17.3 GB42 tok/s16K ctx
dense
F
F Tier -- Can't Run (42 models)
Too heavy for this hardware.
Qwen2.5 1.5B Instruct1.5B16.1 GB21 tok/s11386K ctx
dense
DeepSeek R1 Distill Qwen 1.5B1.5B16.1 GB21 tok/s11386K ctx
dense
Yi Coder 1.5B Chat
D401.5B16.1 GB21 tok/s11386K ctx
dense
stablelm 2 zephyr 1.6b
D391.6B16.2 GB22 tok/s10669K ctx
dense
logos16v2 stablelm2 1.6b i1
D391.6B16.2 GB22 tok/s10669K ctx
dense
Yi Coder 1.5B
D391.5B16.1 GB21 tok/s11386K ctx
dense
Falcon H1 1.5B Instruct1.5B16.1 GB21 tok/s11386K ctx
dense
TinyLlama 1.1B Chat v1.0
D391.1B15.8 GB15 tok/s15556K ctx
dense
Qwen 2.5 0.5B0.5B15.5 GB7 tok/s131K ctx
dense
EXAONE 4.0 1.2B
D391.2B15.9 GB17 tok/s14253K ctx
dense
Llama 3.2 1B Instruct Q8 0
D391B15.9 GB14 tok/s17091K ctx
dense
TinyLlama 1.1B Chat v0.3
D391.1B15.8 GB15 tok/s15556K ctx
dense
TinyLlama 1.1B Chat v0.61.1B15.8 GB15 tok/s15556K ctx
dense
Llama 3.2 1B Instruct
D381B15.7 GB14 tok/s17120K ctx
dense
gemma 3 1b it
D381B15.7 GB14 tok/s17120K ctx
dense
TinyLlama 1.1B Chat v1.0 imatrix
D381.1B15.8 GB15 tok/s15556K ctx
dense
Falcon3 1B Instruct abliterated
D381B15.7 GB14 tok/s17120K ctx
dense
SmolVLM 500M Instruct
D360.5B15.5 GB7 tok/s20094K ctx
dense
embeddinggemma 300M
D350.3B15.3 GB4 tok/s20121K ctx
dense
granite embedding 107m multilingual
D330.11B15.2 GB2 tok/s20150K ctx
dense
Falcon H1 Tiny 90M Instruct
D330.09B15.2 GB2 tok/s20151K ctx
dense
Qwen3-Coder 480B A35B Instruct480B310.7 GB6 tok/s4K ctx
moe
Qwen 3.5 397B A17B397B260.0 GB12 tok/s4K ctx
moe
Mistral Large 3675B432.1 GB5 tok/s4K ctx
moe
GLM-5744B487.9 GB4 tok/s4K ctx
moe
GLM-5.1754B494.0 GB4 tok/s4K ctx
moe
Kimi K2.51000B632.4 GB3 tok/s4K ctx
moe
Kimi K2.61000B632.4 GB3 tok/s4K ctx
moe
DeepSeek V3 671B671B483.9 GB4 tok/s4K ctx
moe
DeepSeek V3.1 671B671B483.9 GB4 tok/s4K ctx
moe
DeepSeek V4 Pro1600B878.9 GB2 tok/s4K ctx
moe
DeepSeek V4 Flash284B174.3 GB43 tok/s4K ctx
moe
DeepSeek R1 671B671B483.9 GB4 tok/s4K ctx
moe
DeepSeek Coder V2 236B236B217.6 GB24 tok/s4K ctx
moe
DeepSeek V2.5 236B236B217.6 GB24 tok/s4K ctx
moe
Llama 4 Maverick 17B 128E400B261.9 GB12 tok/s4K ctx
moe
Llama 3.1 405B405B269.7 GB4 tok/s4K ctx
dense
DeepSeek V3.2671B424.8 GB5 tok/s4K ctx
moe
K EXAONE 236B A23B
F0236B186.6 GB12 tok/s4K ctx
dense
Baichuan M3 235B
F0235B185.9 GB12 tok/s4K ctx
dense
Baichuan M3 235B i1
F0235B185.9 GB12 tok/s4K ctx
dense
Qwen3.5 397B A17B397B303.7 GB3 tok/s4K ctx
dense
Compare with...
See How Other Hardware Stacks Up
Check tier lists for similar GPUs and Apple Silicon configs