AI Compatibility Tier List
NVIDIA B200 180GB
374 AI models ranked by how well they run on your 180 GB VRAM hardware. Best match: Devstral 2 123B Instruct (score: 97).
374
Models tested
336
Can run
180 GB
VRAM
97
Best tok/s
Tier Distribution
S 27
A 54
B 64
C 191
F 38
S: 27A: 54B: 64C: 191F: 38
S
S Tier -- Perfect Fit (27 models)
These models run flawlessly with headroom to spare.
Devstral 2 123B Instruct123B99.3 GB97 tok/s256K ctx
dense
DeepSeek V4 Flash284B178.2 GB145 tok/s38K ctx
moe
Qwen 3.5 122B A10B122B95.8 GB270 tok/s131K ctx
moe
Mistral Small 4 119B119B96.9 GB293 tok/s256K ctx
moe
GPT-OSS 120B117B95.2 GB102 tok/s131K ctx
dense
Pixtral Large 124B124B99.9 GB97 tok/s131K ctx
dense
Command A 111B111B90.5 GB108 tok/s262K ctx
dense
Qwen 3 235B A22B235B165.1 GB137 tok/s99K ctx
moe
Qwen 2.5 VL 72B72B67.7 GB166 tok/s33K ctx
dense
MiniMax M2.7
S90230B163.0 GB156 tok/s88K ctx
moe
Qwen3-Coder-Next80B69.2 GB454 tok/s256K ctx
moe
Leanstral 119B A6B119B100.3 GB269 tok/s161K ctx
moe
Qwen3-Coder 30B A3B Instruct30.5B39.0 GB1016 tok/s256K ctx
moe
Qwen 3.6 35B A3B35B44.4 GB854 tok/s262K ctx
moe
Qwen 3.5 27B27B38.5 GB378 tok/s131K ctx
dense
Qwen3-VL 30B A3B Instruct30B38.7 GB1051 tok/s256K ctx
moe
Qwen 3.6 27B27B36.3 GB275 tok/s262K ctx
dense
Qwen 3.5 35B A3B35B41.7 GB929 tok/s131K ctx
moe
Qwen 3 32B32B42.3 GB374 tok/s131K ctx
dense
Magistral Small 250724B36.0 GB336 tok/s131K ctx
dense
Devstral Small 2 24B Instruct24B36.0 GB336 tok/s256K ctx
dense
Qwen 3 30B A3B30.5B39.0 GB1016 tok/s131K ctx
moe
Nemotron 3 Nano 30B30B39.6 GB395 tok/s131K ctx
dense
Qwen 3.5 9B9B26.6 GB126 tok/s131K ctx
dense
Qwen 3 14B14B29.9 GB196 tok/s131K ctx
dense
Devstral Small 1.124B36.0 GB336 tok/s131K ctx
dense
Phi-4-reasoning-plus 14B14.7B30.9 GB206 tok/s33K ctx
dense
A
A Tier -- Great Fit (54 models)
Excellent performance with comfortable memory margins.
Gemma 4 31B30.7B52.3 GB234 tok/s156K ctx
dense
Qwen 3 8B8B26.0 GB112 tok/s131K ctx
dense
GPT-OSS 20B21B34.2 GB1290 tok/s128K ctx
moe
Nemotron Cascade 2 30B A3B30B40.1 GB1039 tok/s262K ctx
moe
Llama 3.3 70B70B66.5 GB171 tok/s128K ctx
dense
Qwen 3.5 4B4B23.5 GB56 tok/s131K ctx
dense
EXAONE 4.0 32B32B42.3 GB372 tok/s131K ctx
dense
Gemma 4 26B A4B25.2B37.9 GB1091 tok/s256K ctx
moe
Qwen 2.5 72B72B67.7 GB166 tok/s131K ctx
dense
Llama 3.1 70B70B66.5 GB171 tok/s128K ctx
dense
Ministral 3 14B14B29.9 GB196 tok/s262K ctx
multimodal
Qwen 2.5 32B32B42.3 GB372 tok/s131K ctx
dense
Codestral 2 25.0822B34.8 GB308 tok/s256K ctx
dense
Llama 4 Scout 17B 16E109B88.3 GB257 tok/s517K ctx
moe
Mistral Small 3.2 24B24B36.0 GB336 tok/s131K ctx
vision
Kimi Linear 48B A3B48B49.1 GB230 tok/s1049K ctx
linear
Nemotron Nano 8B8B25.7 GB112 tok/s131K ctx
dense
Gemma 3 27B27B46.6 GB266 tok/s131K ctx
dense
Phi-4 Mini Reasoning 4B3.8B22.7 GB53 tok/s131K ctx
dense
LFM2 24B24B36.0 GB336 tok/s131K ctx
dense
OLMo 2 32B32B42.3 GB372 tok/s4K ctx
dense
Granite 4.1 30B30B41.1 GB395 tok/s131K ctx
dense
CogVLM2 19B19B32.9 GB266 tok/s8K ctx
dense
Mistral Small 24B24B36.0 GB336 tok/s33K ctx
dense
Mistral Small 3.1 24B24B36.0 GB336 tok/s131K ctx
dense
InternVL2 8B8B25.7 GB112 tok/s8K ctx
dense
Phi-4 14B14B30.5 GB196 tok/s16K ctx
dense
Qwen 2.5 14B14B30.4 GB196 tok/s131K ctx
dense
DeepSeek R1 Distill 70B70B66.5 GB171 tok/s131K ctx
dense
Granite Code 20B20B34.3 GB280 tok/s8K ctx
dense
Ministral 3 8B8B26.0 GB112 tok/s262K ctx
multimodal
MiniCPM-V 2.6 8B8B25.7 GB112 tok/s2K ctx
dense
Gemma 3 12B12B31.1 GB168 tok/s131K ctx
dense
Qwen 2.5 Coder 32B32B42.3 GB372 tok/s131K ctx
dense
Qwen 3 4B4B23.5 GB56 tok/s33K ctx
dense
Nemotron Nano 9B v29B26.8 GB126 tok/s131K ctx
dense
SQLCoder 7B7B25.1 GB98 tok/s8K ctx
dense
Qwen 2.5 VL 7B7B24.0 GB98 tok/s33K ctx
dense
DeepSeek Coder V2 16B16B32.0 GB1639 tok/s131K ctx
moe
Granite Code 34B34B43.3 GB351 tok/s8K ctx
dense
Jina Embeddings v30.57B22.0 GB8 tok/s8K ctx
dense
Command R 35B35B42.7 GB342 tok/s131K ctx
dense
Magistral 7B7B25.1 GB98 tok/s8K ctx
dense
CodeGeeX 4 9B9B25.0 GB126 tok/s131K ctx
dense
BGE M30.57B21.2 GB8 tok/s8K ctx
dense
OLMo 2 13B13B29.3 GB182 tok/s33K ctx
dense
StarCoder 15B15B44.3 GB210 tok/s8K ctx
dense
DeepSeek R1 Distill 32B32B42.3 GB372 tok/s33K ctx
dense
Gemma 4 E4B8B25.1 GB112 tok/s128K ctx
dense
CodeLlama 13B Instruct13B39.0 GB182 tok/s16K ctx
dense
Nemotron 70B70B66.5 GB171 tok/s131K ctx
dense
Qwen 2.5 7B7B24.0 GB98 tok/s131K ctx
dense
MPT-30B-Instruct30B63.9 GB317 tok/s8K ctx
dense
Granite Code 8B8B25.7 GB112 tok/s8K ctx
dense
B
B Tier -- Good Fit (64 models)
Solid performance, may need quantization or reduced context.
DeepSeek R1 Distill 14B14B30.4 GB196 tok/s33K ctx
dense
DevStral 7B7B25.1 GB98 tok/s8K ctx
dense
LLaVA 1.6 13B13B39.0 GB182 tok/s4K ctx
dense
Codestral Mamba 7B7B23.7 GB98 tok/s262K ctx
state-space
Command R+ 104B104B85.8 GB115 tok/s131K ctx
dense
mxbai Embed Large0.34B21.1 GB5 tok/s1K ctx
dense
Granite 4.1 8B8B26.2 GB112 tok/s131K ctx
dense
CodeLlama 7B Instruct7B31.0 GB98 tok/s16K ctx
dense
Pixtral 12B12B28.7 GB168 tok/s131K ctx
dense
StarCoder 7B7B30.5 GB98 tok/s8K ctx
dense
Falcon 40B Instruct40B49.5 GB259 tok/s8K ctx
dense
Snowflake Arctic Embed L0.34B21.1 GB5 tok/s1K ctx
dense
Mixtral 8x22B141B108.3 GB162 tok/s66K ctx
moe
Qwen 2.5 Coder 3B3B22.9 GB42 tok/s131K ctx
dense
Nous Dolphin 13B13B40.5 GB182 tok/s16K ctx
dense
WizardLM 13B13B39.0 GB182 tok/s8K ctx
dense
Vicuna 13B13B39.0 GB182 tok/s4K ctx
dense
Gemma 2 27B27B46.6 GB266 tok/s8K ctx
dense
Nous Hermes 1.09B36.6 GB126 tok/s16K ctx
dense
Gemma 4 E2B5.1B22.5 GB71 tok/s128K ctx
dense
BGE Large EN v1.50.34B21.1 GB5 tok/s1K ctx
dense
GLM-4 9B9B25.0 GB126 tok/s128K ctx
dense
InternLM 7B7B31.0 GB98 tok/s8K ctx
dense
InternLM Chat 7B7B31.0 GB98 tok/s8K ctx
dense
Llama 3.1 8B8B25.7 GB112 tok/s128K ctx
dense
Solar 7B7B26.1 GB98 tok/s8K ctx
dense
WizardMath 7B7B25.1 GB98 tok/s4K ctx
dense
OLMo 2 7B7B25.1 GB98 tok/s4K ctx
dense
Nomic Embed Text v1.50.14B19.7 GB2 tok/s8K ctx
dense
Qwen 2.5 Coder 7B7B24.0 GB98 tok/s131K ctx
dense
Ministral 3 3B3B21.5 GB42 tok/s262K ctx
multimodal
Gemma 3 4B4B23.4 GB56 tok/s128K ctx
dense
Phi 4 Mini 4B4B22.8 GB56 tok/s128K ctx
dense
LLaVA 1.5 7B7B31.0 GB98 tok/s4K ctx
dense
Qwen 2.5 Math 72B72B67.7 GB166 tok/s4K ctx
dense
Mixtral 8x7B47B49.5 GB483 tok/s33K ctx
moe
Qwen 3.5 2B2B21.8 GB28 tok/s131K ctx
dense
Baichuan 13B13B40.5 GB182 tok/s8K ctx
dense
Cerebras-GPT 13B13B38.0 GB182 tok/s131K ctx
dense
DeepSeek R1 Distill 7B7B24.0 GB98 tok/s33K ctx
dense
MPT-7B-Instruct7B31.0 GB98 tok/s8K ctx
dense
Baichuan 7B7B31.0 GB98 tok/s8K ctx
dense
DeepSeek R1 Distill 8B8B25.7 GB112 tok/s33K ctx
dense
Samantha 7B7B25.1 GB98 tok/s4K ctx
dense
Qwen 2.5 3B3B22.9 GB42 tok/s131K ctx
dense
Falcon 7B Instruct7B23.3 GB98 tok/s8K ctx
dense
Phi 3 Mini 3.8B3.8B27.1 GB53 tok/s128K ctx
dense
Qwen 3 1.7B1.7B21.6 GB24 tok/s33K ctx
dense
Gemma 2 9B9B29.5 GB126 tok/s8K ctx
dense
Qwen 2.5 Coder 14B14B30.4 GB196 tok/s131K ctx
dense
All MiniLM L6 v20.02B19.2 GB2 tok/s0K ctx
dense
Yi 1.5 34B34B43.3 GB352 tok/s4K ctx
dense
Llama 3.2 11B Vision11B27.6 GB154 tok/s16K ctx
vision
DeepSeek LLM 67B67B65.6 GB179 tok/s4K ctx
dense
Phi 3.5 Mini 4B4B27.2 GB56 tok/s128K ctx
dense
Granite 4.1 3B3B22.0 GB42 tok/s131K ctx
dense
Qwen 2.5 Coder 1.5B1.5B20.2 GB21 tok/s33K ctx
dense
Granite Code 3B3B23.2 GB42 tok/s8K ctx
dense
Mistral Nemo 12B12B28.7 GB168 tok/s128K ctx
dense
Yi Coder 9B9B25.9 GB126 tok/s131K ctx
dense
Mistral 7B Instruct v0.37B25.1 GB98 tok/s8K ctx
dense
Phi 3 Medium 14B14B30.5 GB196 tok/s128K ctx
dense
Codestral 22B22B34.8 GB308 tok/s33K ctx
dense
InternLM 20B20B51.6 GB280 tok/s8K ctx
dense
C
C Tier -- Tight Fit (191 models)
Runs with offloading or significant compromises.
Llama 3.2 3B3B22.4 GB42 tok/s128K ctx
dense
Ministral 8B8B26.0 GB112 tok/s131K ctx
dense
Qwen3.5 122B A10B122B93.0 GB105 tok/s113K ctx
dense
Solar Open 100B
C53100B91.6 GB110 tok/s137K ctx
dense
Solar Open 100B i1
C53100B91.6 GB110 tok/s137K ctx
dense
Aya Expanse 32B32B40.9 GB374 tok/s8K ctx
dense
Llama 3.3 70B Instruct
C5070B69.8 GB157 tok/s231K ctx
dense
DeepSeek R1 1.5B1.5B20.2 GB21 tok/s33K ctx
dense
TinyLlama 1.1B1.1B19.9 GB15 tok/s4K ctx
dense
Solar Open 69B REAP i1
C5069B69.1 GB160 tok/s235K ctx
dense
SmolLM3 3B3B22.7 GB42 tok/s128K ctx
dense
BaichuanMed OCR 72B i1
C5072B71.3 GB153 tok/s222K ctx
dense
stabilityai japanese stablelm instruct beta 70b
C5070B69.8 GB157 tok/s231K ctx
dense
Granite 3.1 8B8B25.7 GB112 tok/s128K ctx
state-space
Yi 1.5 9B9B25.9 GB126 tok/s4K ctx
dense
Qwen 2.5 Math 7B7B24.0 GB98 tok/s4K ctx
dense
Yi 34B Chat34B43.3 GB352 tok/s200K ctx
dense
Qwen3 48B A4B Savant Commander Distill 12X Closed Open Heretic Uncensored
C4848B53.8 GB230 tok/s375K ctx
dense
Qwen 2.5 1.5B1.5B20.2 GB21 tok/s131K ctx
dense
Qwen3.5 35B A3B35B44.4 GB315 tok/s545K ctx
dense
Qwen3.5 35B A3B
C4735B44.4 GB315 tok/s545K ctx
dense
Hermes 4.3 36B
C4736B45.1 GB306 tok/s528K ctx
dense
StarCoder2 15B15B30.9 GB210 tok/s16K ctx
dense
OpenChat 7B7B25.1 GB98 tok/s8K ctx
dense
Qwen3.5 27B27B38.5 GB378 tok/s731K ctx
dense
gemma 3 27b it27B38.5 GB378 tok/s731K ctx
dense
EXAONE 4.0 32B
C4732B42.2 GB344 tok/s604K ctx
dense
aya expanse 32b heretic MPOA i1
C4732B42.2 GB344 tok/s604K ctx
dense
Baichuan M2 32B Q4 K M
C4732B42.2 GB344 tok/s604K ctx
dense
gemma 3 27b it
C4727B38.5 GB378 tok/s731K ctx
dense
baichuan inc Baichuan M2 32B
C4632B42.2 GB344 tok/s604K ctx
dense
Aya Expanse 8B8B25.7 GB112 tok/s8K ctx
dense
StableLM 2 12B12B39.7 GB168 tok/s4K ctx
dense
Mistral Small 3.2 24B Instruct 250624B36.4 GB336 tok/s833K ctx
dense
cognitivecomputations Dolphin Mistral 24B Venice Edition
C4624B36.4 GB336 tok/s833K ctx
dense
Mistral Small 24B Instruct 2501
C4624B36.4 GB336 tok/s833K ctx
dense
Gemma 3 1B1B19.9 GB14 tok/s33K ctx
dense
mistral small 3.1 24b instruct 2503 hf
C4624B36.4 GB336 tok/s833K ctx
dense
cognitivecomputations Dolphin3.0 R1 Mistral 24B
C4624B36.4 GB336 tok/s833K ctx
dense
Vicuna 7B7B31.0 GB98 tok/s4K ctx
dense
Dolphin Mistral GLM 4.7 Flash 24B Venice Edition Thinking Uncensored i1
C4624B36.4 GB336 tok/s833K ctx
dense
Codestral 22B v0.1
C4622B34.9 GB308 tok/s917K ctx
dense
Codestral 22B v0.1
C4622B34.9 GB308 tok/s917K ctx
dense
cognitivecomputations Dolphin Mistral 24B Venice Edition
C4624B36.4 GB336 tok/s833K ctx
dense
Codestral 22B v0.1
C4622B34.9 GB308 tok/s917K ctx
dense
OpenHermes 2.5 7B7B25.1 GB98 tok/s8K ctx
dense
Codestral 22B v0.1 IMat
C4622B34.9 GB308 tok/s917K ctx
dense
Dolphin 2.9 8B8B25.7 GB112 tok/s33K ctx
dense
Codestral 22B v0.1 i1
C4522B34.9 GB308 tok/s917K ctx
dense
DeepSeek R1 Distill Qwen 14B14B29.1 GB196 tok/s1488K ctx
dense
Codestral 21B Pruned i1
C4521B34.2 GB294 tok/s964K ctx
dense
internlm2 5 20b chat
C4520B33.4 GB280 tok/s1016K ctx
dense
Zephyr 7B Beta7B25.1 GB98 tok/s33K ctx
dense
internlm2 limarp chat 20b
C4520B33.4 GB280 tok/s1016K ctx
dense
internlm2 math plus 20b i1
C4520B33.4 GB280 tok/s1016K ctx
dense
Gemma 2 2B2B21.7 GB28 tok/s8K ctx
dense
Phi 4 reasoning vision 15B
C4515B29.8 GB210 tok/s1383K ctx
dense
StarCoder2 15B
C4515B29.8 GB210 tok/s1383K ctx
dense
Qwen3.5 9B9B25.4 GB126 tok/s2361K ctx
dense
gemma 3 12b it
C4512B27.6 GB168 tok/s1750K ctx
dense
Qwen3.5 9B Uncensored HauhauCS Aggressive
C459B25.4 GB126 tok/s2361K ctx
dense
Codestral RAG 19B Pruned i1
C4519B32.7 GB266 tok/s1074K ctx
dense
NousResearch Hermes 4 14B
C4514B29.1 GB196 tok/s1488K ctx
dense
Starling LM 7B7B25.1 GB98 tok/s8K ctx
dense
Meta Llama 3.1 8B Instruct
C458B24.7 GB112 tok/s2666K ctx
dense
Qwen3.5 9B
C459B25.4 GB126 tok/s2361K ctx
dense
llava llama 3 8b v1 1
C458B24.7 GB112 tok/s2666K ctx
dense
starcoder2 15b instruct v0.1
C4515B29.8 GB210 tok/s1383K ctx
dense
starcoder2 15b i1
C4515B29.8 GB210 tok/s1383K ctx
dense
DeepSeek R1 0528 Qwen3 8B8B24.7 GB112 tok/s2666K ctx
dense
starcoder2 15b instruct v0.1
C4515B29.8 GB210 tok/s1383K ctx
dense
Llama 2 7B Chat
C457B24.0 GB98 tok/s3059K ctx
dense
Meta Llama 3 8B Instruct
C458B24.7 GB112 tok/s2666K ctx
dense
Qwen 2.5 Coder 0.5B0.5B19.4 GB7 tok/s131K ctx
dense
DeepSeek R1 Distill Llama 8B8B24.7 GB112 tok/s2666K ctx
dense
GGUF SOLARized GraniStral 14B 2102 YeAM HCT 32QKV
C4514B29.1 GB196 tok/s1488K ctx
dense
Mistral 7B Instruct v0.2
C457B24.0 GB98 tok/s3059K ctx
dense
SOLAR 10.7B Instruct v1.0 uncensored
C4510.7B26.7 GB150 tok/s1972K ctx
dense
Dolphin3.0 Llama3.1 8B
C458B24.7 GB112 tok/s2666K ctx
dense
Llama 3 8B Instruct 32k v0.1
C458B24.7 GB112 tok/s2666K ctx
dense
Meta Llama 3.1 8B Instruct
C458B24.7 GB112 tok/s2666K ctx
dense
Nous Hermes 2 SOLAR 10.7B
C4510.7B26.7 GB150 tok/s1972K ctx
dense
vntl llama3 8b v2
C458B24.7 GB112 tok/s2666K ctx
dense
HelpingAI 15B i1
C4515B29.8 GB210 tok/s1383K ctx
dense
Mistral 7B Instruct v0.3
C457B24.0 GB98 tok/s3059K ctx
dense
Yi Coder 9B Chat
C459B25.4 GB126 tok/s2361K ctx
dense
internlm JanusCoder 14B
C4514B29.1 GB196 tok/s1488K ctx
dense
DeepSeek R1 0528 Qwen3 8B
C458B24.7 GB112 tok/s2666K ctx
dense
glm 4 9b chat 1m
C459B25.4 GB126 tok/s2361K ctx
dense
GGUF SOLARized GraniStral 14B 1902 YeAM HCT
C4514B29.1 GB196 tok/s1488K ctx
dense
DeepSeek R1 0528 Qwen3 8B
C458B24.7 GB112 tok/s2666K ctx
dense
Qwen3 8B DeepSeek v3.2 Speciale Distill
C458B24.7 GB112 tok/s2666K ctx
dense
Hermes 3 Llama 3.1 8B
C458B24.7 GB112 tok/s2666K ctx
dense
Yi 1.5 9B Chat
C459B25.4 GB126 tok/s2361K ctx
dense
Hermes 2 Pro Llama 3 8B
C458B24.7 GB112 tok/s2666K ctx
dense
Dolphin3.0 Llama3.1 8B
C458B24.7 GB112 tok/s2666K ctx
dense
Mistral 7B Instruct v0.3
C457B24.0 GB98 tok/s3059K ctx
dense
granite 8b code instruct 4k
C458B24.7 GB112 tok/s2666K ctx
dense
Hermes 2 Pro Mistral 7B
C457B24.0 GB98 tok/s3059K ctx
dense
dolphin 2.9.4 llama3.1 8b
C458B24.7 GB112 tok/s2666K ctx
dense
solar finalised finetuned Model 10.7B i1
C4510.7B26.7 GB150 tok/s1972K ctx
dense
zephyr 7B beta
C457B24.0 GB98 tok/s3059K ctx
dense
Nous Hermes 2 Mistral 7B DPO
C457B24.0 GB98 tok/s3059K ctx
dense
SOLAR 10.7B v1.0
C4510.7B26.7 GB150 tok/s1972K ctx
dense
openchat 3.6 8b 20240522 IMat
C458B24.7 GB112 tok/s2666K ctx
dense
Falcon H1R 7B
C457B24.0 GB98 tok/s3059K ctx
dense
aya expanse 8b
C458B24.7 GB112 tok/s2666K ctx
dense
EXAONE 3.5 7.8B Instruct
C457.8B24.6 GB109 tok/s2737K ctx
dense
falcon mamba 7b instruct Q4 K M
C457B24.0 GB98 tok/s3059K ctx
dense
dolphin v2 8b abliterated i1
C458B24.7 GB112 tok/s2666K ctx
dense
zephyr 7B alpha
C457B24.0 GB98 tok/s3059K ctx
dense
EXAONE 3.5 7.8B Instruct
C457.8B24.6 GB109 tok/s2737K ctx
dense
speechless zephyr code functionary 7b
C457B24.0 GB98 tok/s3059K ctx
dense
Yi 9B Coder i1
C459B25.4 GB126 tok/s2361K ctx
dense
EXAONE 3.5 7.8B Instruct i1
C457.8B24.6 GB109 tok/s2737K ctx
dense
StarCoder2 7B
C457B24.0 GB98 tok/s3059K ctx
dense
aya expanse 8b orthogonal heretic i1
C458B24.7 GB112 tok/s2666K ctx
dense
exaone 3.0 7.8b it
C457.8B24.6 GB109 tok/s2737K ctx
dense
japanese stablelm instruct gamma 7B
C447B24.0 GB98 tok/s3059K ctx
dense
HelpingAI2 9B
C449B25.4 GB126 tok/s2361K ctx
dense
HelpingAI2.5 10B i1
C4410B26.2 GB140 tok/s2116K ctx
dense
Falcon H1 7B Instruct
C447B24.0 GB98 tok/s3059K ctx
dense
Falcon H1R 7B7B24.0 GB98 tok/s3059K ctx
dense
Yi 1.5 6B Chat
C446B23.3 GB84 tok/s3583K ctx
dense
Helply 10.2b chat i1
C4410.2B26.3 GB143 tok/s2073K ctx
dense
aya expanse 8b orthogonal heretic
C448B24.7 GB112 tok/s2666K ctx
dense
HelpingAI2 9B i1
C449B25.4 GB126 tok/s2361K ctx
dense
HelpingAI 9B 200k i1
C449B25.4 GB126 tok/s2361K ctx
dense
stablelm 2 zephyr 1 6b
C446B23.3 GB84 tok/s3583K ctx
dense
HelpingAI 9B i1
C449B25.4 GB126 tok/s2361K ctx
dense
Mamba Codestral 7B v0.1
C447B24.0 GB98 tok/s3059K ctx
dense
stabilityai japanese stablelm base gamma 7b
C447B24.0 GB98 tok/s3059K ctx
dense
baichuan2 7b chat
C447B24.0 GB98 tok/s3059K ctx
dense
internlm2 5 1 8b chat i1
C448B24.7 GB112 tok/s2666K ctx
dense
DiscoPOP zephyr 7b gemma
C447B24.0 GB98 tok/s3059K ctx
dense
internlm3 8b instruct abliterated i1
C448B24.7 GB112 tok/s2666K ctx
dense
starcoder2 7b
C447B24.0 GB98 tok/s3059K ctx
dense
zephyr 7b beta Mistral 7B Instruct v0.2
C447B24.0 GB98 tok/s3059K ctx
dense
OpenChat 3.5 7B Qwen v2.0 i1
C447B24.0 GB98 tok/s3059K ctx
dense
OpenChat 3.5 7B Starling v2.0 i1
C447B24.0 GB98 tok/s3059K ctx
dense
internlm2 math plus 7b IMat
C447B24.0 GB98 tok/s3059K ctx
dense
CodeNinja 1.0 OpenChat 7B i1
C447B24.0 GB98 tok/s3059K ctx
dense
internlm2 5 7b chat i1
C447B24.0 GB98 tok/s3059K ctx
dense
OpenSafetyLab MD Judge v0 2 internlm2 7b
C447B24.0 GB98 tok/s3059K ctx
dense
MD Judge v0 2 internlm2 7b i1
C447B24.0 GB98 tok/s3059K ctx
dense
Yi 1.5 6B Chat
C446B23.3 GB84 tok/s3583K ctx
dense
Yi 1.5 6B6B23.5 GB84 tok/s4K ctx
dense
zephyr 7b gemma sft african ultrachat 100k
C447B24.0 GB98 tok/s3059K ctx
dense
stablelm 2 1 6b chat imatrix
C446B23.3 GB84 tok/s3583K ctx
dense
jointpreferences mistral 7b sft helpful
C447B24.0 GB98 tok/s3059K ctx
dense
zephyr 7b dpo full i1
C447B24.0 GB98 tok/s3059K ctx
dense
blossom v3 baichuan2 7b i1
C447B24.0 GB98 tok/s3059K ctx
dense
blossom v1 baichuan 7b i1
C447B24.0 GB98 tok/s3059K ctx
dense
Neural Chat 7B7B25.1 GB98 tok/s8K ctx
dense
HelpingAI2 6B
C446B23.3 GB84 tok/s3583K ctx
dense
StarCoder2 7B7B23.7 GB98 tok/s16K ctx
dense
HelpingAI2 6B i1
C446B23.3 GB84 tok/s3583K ctx
dense
DeepSeek LLM 7B7B30.5 GB98 tok/s4K ctx
dense
Nemotron Mini 4B4B23.3 GB56 tok/s4K ctx
dense
HelpingAI2.5 5B i1
C435B22.5 GB70 tok/s4316K ctx
dense
Qwen3.5 4B4B21.8 GB56 tok/s5416K ctx
dense
gemma 3 4b it
C434B21.8 GB56 tok/s5416K ctx
dense
gemma 3 4b it
C434B21.8 GB56 tok/s5416K ctx
dense
Qwen 3 0.6B0.6B20.1 GB8 tok/s33K ctx
dense
Llama 3.2 1B1B20.0 GB14 tok/s128K ctx
dense
Llama 3.2 3B Instruct
C423B21.4 GB42 tok/s7234K ctx
dense
Qwen2.5 3B Instruct3B21.1 GB42 tok/s7249K ctx
dense
Llama 3.2 3B Instruct
C423B21.1 GB42 tok/s7249K ctx
dense
Ministral 3 3B Instruct 2512
C423B21.1 GB42 tok/s7249K ctx
dense
HELVETE 3B
C423B21.1 GB42 tok/s7249K ctx
dense
Hermes 3 Llama 3.2 3B
C423B21.1 GB42 tok/s7249K ctx
dense
stablelm zephyr 3b
C423B21.1 GB42 tok/s7249K ctx
dense
StarCoder2 3B
C423B21.1 GB42 tok/s7249K ctx
dense
AI21 Jamba Reasoning 3B
C423B21.1 GB42 tok/s7249K ctx
dense
stablelm 3b 4e1t
C413B21.1 GB42 tok/s7249K ctx
dense
ai21labs AI21 Jamba Reasoning 3B
C413B21.1 GB42 tok/s7249K ctx
dense
ai21labs AI21 Jamba2 3B
C413B21.1 GB42 tok/s7249K ctx
dense
HelpingAI 3B hindi i1
C413B21.1 GB42 tok/s7249K ctx
dense
AI21 Jamba2 3B
C413B21.1 GB42 tok/s7249K ctx
dense
HelpingAI 3B hindi
C413B21.1 GB42 tok/s7249K ctx
dense
AI21 Jamba2 3B i1
C413B21.1 GB42 tok/s7249K ctx
dense
EXAONE 3.5 2.4B Instruct
C412.4B20.6 GB34 tok/s9082K ctx
dense
gemma 2 2b it
C412B20.8 GB28 tok/s10886K ctx
dense
gemma 2b2B20.4 GB28 tok/s10914K ctx
dense
Gemmasutra Mini 2B v1
C412B20.4 GB28 tok/s10914K ctx
dense
K EXAONE 236B A23B
C41236B190.5 GB36 tok/s10K ctx
dense
Qwen 3.5 0.6B0.6B20.1 GB8 tok/s131K ctx
dense
gemma 2 2b it
C412B20.4 GB28 tok/s10914K ctx
dense
Baichuan M3 235B
C41235B189.8 GB37 tok/s10K ctx
dense
Baichuan M3 235B i1
C41235B189.8 GB37 tok/s10K ctx
dense
StarCoder2 3B3B21.2 GB42 tok/s16K ctx
dense
F
F Tier -- Can't Run (38 models)
Too heavy for this hardware.
Qwen2.5 1.5B Instruct1.5B20.0 GB21 tok/s14580K ctx
dense
DeepSeek R1 Distill Qwen 1.5B1.5B20.0 GB21 tok/s14580K ctx
dense
Yi Coder 1.5B Chat
D401.5B20.0 GB21 tok/s14580K ctx
dense
stablelm 2 zephyr 1.6b
D391.6B20.1 GB22 tok/s13664K ctx
dense
logos16v2 stablelm2 1.6b i1
D391.6B20.1 GB22 tok/s13664K ctx
dense
Yi Coder 1.5B
D391.5B20.0 GB21 tok/s14580K ctx
dense
Falcon H1 1.5B Instruct1.5B20.0 GB21 tok/s14580K ctx
dense
TinyLlama 1.1B Chat v1.0
D391.1B19.7 GB15 tok/s19913K ctx
dense
Qwen 2.5 0.5B0.5B19.4 GB7 tok/s131K ctx
dense
EXAONE 4.0 1.2B
D391.2B19.8 GB17 tok/s18246K ctx
dense
Llama 3.2 1B Instruct Q8 0
D391B19.8 GB14 tok/s21884K ctx
dense
TinyLlama 1.1B Chat v0.3
D381.1B19.7 GB15 tok/s19913K ctx
dense
TinyLlama 1.1B Chat v0.61.1B19.7 GB15 tok/s19913K ctx
dense
Llama 3.2 1B Instruct
D381B19.6 GB14 tok/s21912K ctx
dense
gemma 3 1b it
D381B19.6 GB14 tok/s21912K ctx
dense
TinyLlama 1.1B Chat v1.0 imatrix
D381.1B19.7 GB15 tok/s19913K ctx
dense
Falcon3 1B Instruct abliterated
D381B19.6 GB14 tok/s21912K ctx
dense
SmolVLM 500M Instruct
D360.5B19.4 GB7 tok/s25710K ctx
dense
embeddinggemma 300M
D350.3B19.2 GB4 tok/s25737K ctx
dense
granite embedding 107m multilingual
D330.11B19.1 GB2 tok/s25766K ctx
dense
Falcon H1 Tiny 90M Instruct
D330.09B19.1 GB2 tok/s25767K ctx
dense
Qwen3-Coder 480B A35B Instruct480B314.6 GB24 tok/s4K ctx
moe
Qwen 3.5 397B A17B397B263.9 GB41 tok/s4K ctx
moe
Mistral Large 3675B436.0 GB10 tok/s4K ctx
moe
GLM-5744B491.8 GB7 tok/s4K ctx
moe
GLM-5.1754B497.9 GB7 tok/s4K ctx
moe
Kimi K2.51000B636.3 GB6 tok/s4K ctx
moe
Kimi K2.61000B636.3 GB6 tok/s4K ctx
moe
DeepSeek V3 671B671B487.8 GB8 tok/s4K ctx
moe
DeepSeek V3.1 671B671B487.8 GB8 tok/s4K ctx
moe
DeepSeek V4 Pro1600B882.8 GB4 tok/s4K ctx
moe
DeepSeek R1 671B671B487.8 GB8 tok/s4K ctx
moe
DeepSeek Coder V2 236B236B221.5 GB76 tok/s5K ctx
moe
DeepSeek V2.5 236B236B221.5 GB76 tok/s5K ctx
moe
Llama 4 Maverick 17B 128E400B265.8 GB40 tok/s4K ctx
moe
Llama 3.1 405B405B273.6 GB13 tok/s4K ctx
dense
DeepSeek V3.2671B428.7 GB11 tok/s4K ctx
moe
Qwen3.5 397B A17B397B307.6 GB10 tok/s4K ctx
dense
Compare with...
See How Other Hardware Stacks Up
Check tier lists for similar GPUs and Apple Silicon configs