AI Compatibility Tier List
RTX 3080 10GB
374 AI models ranked by how well they run on your 10 GB VRAM hardware. Best match: Qwen 3.5 9B (score: 95).
374
Models tested
208
Can run
10 GB
VRAM
91
Best tok/s
Tier Distribution
S 9
A 33
B 86
C 80
F 166
S: 9A: 33B: 86C: 80F: 166
S
S Tier -- Perfect Fit (9 models)
These models run flawlessly with headroom to spare.
Qwen 3.5 9B9B9.6 GB91 tok/s19K ctx
dense
Qwen 3 8B8B9.0 GB96 tok/s23K ctx
dense
Qwen 3.5 4B4B6.5 GB48 tok/s41K ctx
dense
Nemotron Nano 8B8B8.7 GB96 tok/s26K ctx
dense
Phi-4 Mini Reasoning 4B3.8B5.7 GB46 tok/s63K ctx
dense
Qwen 2.5 VL 7B7B7.0 GB84 tok/s33K ctx
dense
SQLCoder 7B7B8.1 GB84 tok/s8K ctx
dense
InternVL2 8B8B8.7 GB96 tok/s8K ctx
dense
Qwen 3 4B4B6.5 GB48 tok/s33K ctx
dense
A
A Tier -- Great Fit (33 models)
Excellent performance with comfortable memory margins.
CodeGeeX 4 9B9B8.0 GB97 tok/s68K ctx
dense
Magistral 7B7B8.1 GB84 tok/s8K ctx
dense
Ministral 3 8B8B9.0 GB96 tok/s23K ctx
multimodal
MiniCPM-V 2.6 8B8B8.7 GB96 tok/s2K ctx
dense
Gemma 4 E4B8B8.1 GB81 tok/s40K ctx
dense
Nemotron Nano 9B v29B9.8 GB95 tok/s17K ctx
dense
Qwen 2.5 7B7B7.0 GB84 tok/s72K ctx
dense
DevStral 7B7B8.1 GB84 tok/s8K ctx
dense
Jina Embeddings v30.57B5.0 GB7 tok/s8K ctx
dense
Codestral Mamba 7B7B6.7 GB84 tok/s126K ctx
state-space
Granite Code 8B8B8.7 GB96 tok/s8K ctx
dense
BGE M30.57B4.2 GB7 tok/s8K ctx
dense
GLM-4 9B9B8.0 GB97 tok/s68K ctx
dense
Granite 4.1 8B8B9.2 GB96 tok/s21K ctx
dense
WizardMath 7B7B8.1 GB84 tok/s4K ctx
dense
OLMo 2 7B7B8.1 GB84 tok/s4K ctx
dense
Qwen 2.5 Coder 3B3B5.9 GB36 tok/s46K ctx
dense
Qwen 2.5 Coder 7B7B7.0 GB84 tok/s72K ctx
dense
Gemma 4 E2B5.1B5.5 GB61 tok/s128K ctx
dense
mxbai Embed Large0.34B4.1 GB4 tok/s1K ctx
dense
Llama 3.1 8B8B8.7 GB96 tok/s26K ctx
dense
Gemma 3 4B4B6.4 GB48 tok/s44K ctx
dense
Solar 7B7B9.1 GB84 tok/s8K ctx
dense
Snowflake Arctic Embed L0.34B4.1 GB4 tok/s1K ctx
dense
DeepSeek R1 Distill 7B7B7.0 GB84 tok/s33K ctx
dense
Phi 4 Mini 4B4B5.8 GB48 tok/s62K ctx
dense
Samantha 7B7B8.1 GB84 tok/s4K ctx
dense
BGE Large EN v1.50.34B4.1 GB4 tok/s1K ctx
dense
Falcon 7B Instruct7B6.3 GB84 tok/s8K ctx
dense
Ministral 3 3B3B4.5 GB36 tok/s137K ctx
multimodal
DeepSeek R1 Distill 8B8B8.7 GB96 tok/s26K ctx
dense
Qwen 2.5 3B3B5.9 GB36 tok/s46K ctx
dense
Qwen 3.5 2B2B4.8 GB24 tok/s64K ctx
dense
B
B Tier -- Good Fit (86 models)
Solid performance, may need quantization or reduced context.
Mistral 7B Instruct v0.37B8.1 GB84 tok/s8K ctx
dense
Phi 3 Mini 3.8B3.8B10.1 GB46 tok/s16K ctx
dense
Nomic Embed Text v1.50.14B2.7 GB2 tok/s8K ctx
dense
Granite Code 3B3B6.2 GB36 tok/s8K ctx
dense
Phi 3.5 Mini 4B4B10.2 GB48 tok/s15K ctx
dense
Qwen 3 1.7B1.7B4.6 GB20 tok/s33K ctx
dense
Granite 4.1 3B3B5.0 GB36 tok/s82K ctx
dense
Yi Coder 9B9B8.9 GB96 tok/s29K ctx
dense
Pixtral 12B12B11.7 GB45 tok/s5K ctx
dense
Ministral 8B8B9.0 GB96 tok/s23K ctx
dense
Llama 3.2 3B3B5.4 GB36 tok/s59K ctx
dense
Qwen 2.5 Coder 1.5B1.5B3.2 GB18 tok/s33K ctx
dense
All MiniLM L6 v20.02B2.2 GB2 tok/s0K ctx
dense
Qwen 2.5 Math 7B7B7.0 GB84 tok/s4K ctx
dense
SmolLM3 3B3B5.7 GB36 tok/s51K ctx
dense
OpenChat 7B7B8.1 GB84 tok/s8K ctx
dense
Granite 3.1 8B8B8.7 GB96 tok/s26K ctx
state-space
Yi 1.5 9B9B8.9 GB96 tok/s4K ctx
dense
OpenHermes 2.5 7B7B8.1 GB84 tok/s8K ctx
dense
Meta Llama 3.1 8B Instruct
B578B7.7 GB96 tok/s55K ctx
dense
llava llama 3 8b v1 1
B578B7.7 GB96 tok/s55K ctx
dense
Zephyr 7B Beta7B8.1 GB84 tok/s31K ctx
dense
DeepSeek R1 0528 Qwen3 8B8B7.7 GB96 tok/s55K ctx
dense
Meta Llama 3 8B Instruct
B578B7.7 GB96 tok/s55K ctx
dense
DeepSeek R1 Distill Llama 8B8B7.7 GB96 tok/s55K ctx
dense
Dolphin3.0 Llama3.1 8B
B578B7.7 GB96 tok/s55K ctx
dense
Llama 3 8B Instruct 32k v0.1
B578B7.7 GB96 tok/s55K ctx
dense
Meta Llama 3.1 8B Instruct
B578B7.7 GB96 tok/s55K ctx
dense
vntl llama3 8b v2
B578B7.7 GB96 tok/s55K ctx
dense
DeepSeek R1 0528 Qwen3 8B
B578B7.7 GB96 tok/s55K ctx
dense
DeepSeek R1 0528 Qwen3 8B
B578B7.7 GB96 tok/s55K ctx
dense
Llama 2 7B Chat
B577B7.0 GB84 tok/s75K ctx
dense
Llama 3.2 11B Vision11B10.6 GB61 tok/s11K ctx
vision
Qwen3 8B DeepSeek v3.2 Speciale Distill
B578B7.7 GB96 tok/s55K ctx
dense
Starling LM 7B7B8.1 GB84 tok/s8K ctx
dense
Hermes 3 Llama 3.1 8B
B578B7.7 GB96 tok/s55K ctx
dense
Mistral 7B Instruct v0.2
B577B7.0 GB84 tok/s75K ctx
dense
Mistral 7B Instruct v0.3
B577B7.0 GB84 tok/s75K ctx
dense
Hermes 2 Pro Llama 3 8B
B578B7.7 GB96 tok/s55K ctx
dense
Dolphin3.0 Llama3.1 8B
B578B7.7 GB96 tok/s55K ctx
dense
granite 8b code instruct 4k
B578B7.7 GB96 tok/s55K ctx
dense
dolphin 2.9.4 llama3.1 8b
B578B7.7 GB96 tok/s55K ctx
dense
openchat 3.6 8b 20240522 IMat
B578B7.7 GB96 tok/s55K ctx
dense
EXAONE 3.5 7.8B Instruct
B577.8B7.6 GB94 tok/s58K ctx
dense
aya expanse 8b
B578B7.7 GB96 tok/s55K ctx
dense
EXAONE 3.5 7.8B Instruct
B577.8B7.6 GB94 tok/s58K ctx
dense
dolphin v2 8b abliterated i1
B578B7.7 GB96 tok/s55K ctx
dense
Mistral 7B Instruct v0.3
B577B7.0 GB84 tok/s75K ctx
dense
EXAONE 3.5 7.8B Instruct i1
B577.8B7.6 GB94 tok/s58K ctx
dense
Hermes 2 Pro Mistral 7B
B577B7.0 GB84 tok/s75K ctx
dense
exaone 3.0 7.8b it
B577.8B7.6 GB94 tok/s58K ctx
dense
aya expanse 8b orthogonal heretic i1
B578B7.7 GB96 tok/s55K ctx
dense
zephyr 7B beta
B577B7.0 GB84 tok/s75K ctx
dense
Nous Hermes 2 Mistral 7B DPO
B577B7.0 GB84 tok/s75K ctx
dense
Falcon H1R 7B
B567B7.0 GB84 tok/s75K ctx
dense
aya expanse 8b orthogonal heretic
B568B7.7 GB96 tok/s55K ctx
dense
falcon mamba 7b instruct Q4 K M
B567B7.0 GB84 tok/s75K ctx
dense
zephyr 7B alpha
B567B7.0 GB84 tok/s75K ctx
dense
speechless zephyr code functionary 7b
B567B7.0 GB84 tok/s75K ctx
dense
StarCoder2 7B
B567B7.0 GB84 tok/s75K ctx
dense
japanese stablelm instruct gamma 7B
B567B7.0 GB84 tok/s75K ctx
dense
internlm2 5 1 8b chat i1
B568B7.7 GB96 tok/s55K ctx
dense
internlm3 8b instruct abliterated i1
B568B7.7 GB96 tok/s55K ctx
dense
Falcon H1 7B Instruct
B567B7.0 GB84 tok/s75K ctx
dense
Falcon H1R 7B7B7.0 GB84 tok/s75K ctx
dense
Mamba Codestral 7B v0.1
B567B7.0 GB84 tok/s75K ctx
dense
stabilityai japanese stablelm base gamma 7b
B567B7.0 GB84 tok/s75K ctx
dense
baichuan2 7b chat
B567B7.0 GB84 tok/s75K ctx
dense
DiscoPOP zephyr 7b gemma
B567B7.0 GB84 tok/s75K ctx
dense
starcoder2 7b
B567B7.0 GB84 tok/s75K ctx
dense
zephyr 7b beta Mistral 7B Instruct v0.2
B567B7.0 GB84 tok/s75K ctx
dense
OpenChat 3.5 7B Qwen v2.0 i1
B567B7.0 GB84 tok/s75K ctx
dense
OpenChat 3.5 7B Starling v2.0 i1
B567B7.0 GB84 tok/s75K ctx
dense
internlm2 math plus 7b IMat
B567B7.0 GB84 tok/s75K ctx
dense
CodeNinja 1.0 OpenChat 7B i1
B567B7.0 GB84 tok/s75K ctx
dense
internlm2 5 7b chat i1
B567B7.0 GB84 tok/s75K ctx
dense
OpenSafetyLab MD Judge v0 2 internlm2 7b
B567B7.0 GB84 tok/s75K ctx
dense
MD Judge v0 2 internlm2 7b i1
B567B7.0 GB84 tok/s75K ctx
dense
zephyr 7b gemma sft african ultrachat 100k
B567B7.0 GB84 tok/s75K ctx
dense
jointpreferences mistral 7b sft helpful
B567B7.0 GB84 tok/s75K ctx
dense
zephyr 7b dpo full i1
B567B7.0 GB84 tok/s75K ctx
dense
blossom v3 baichuan2 7b i1
B567B7.0 GB84 tok/s75K ctx
dense
blossom v1 baichuan 7b i1
B567B7.0 GB84 tok/s75K ctx
dense
Neural Chat 7B7B8.1 GB84 tok/s8K ctx
dense
Aya Expanse 8B8B8.7 GB96 tok/s8K ctx
dense
StarCoder2 7B7B6.7 GB84 tok/s16K ctx
dense
C
C Tier -- Tight Fit (80 models)
Runs with offloading or significant compromises.
Yi 1.5 6B6B6.5 GB72 tok/s4K ctx
dense
Yi 1.5 6B Chat
C556B6.3 GB72 tok/s101K ctx
dense
stablelm 2 zephyr 1 6b
C556B6.3 GB72 tok/s101K ctx
dense
Dolphin 2.9 8B8B8.7 GB96 tok/s26K ctx
dense
Yi 1.5 6B Chat
C546B6.3 GB72 tok/s101K ctx
dense
stablelm 2 1 6b chat imatrix
C546B6.3 GB72 tok/s101K ctx
dense
DeepSeek R1 1.5B1.5B3.2 GB18 tok/s33K ctx
dense
Qwen3.5 9B9B8.4 GB88 tok/s40K ctx
dense
HelpingAI2 6B
C546B6.3 GB72 tok/s101K ctx
dense
Qwen3.5 9B Uncensored HauhauCS Aggressive
C549B8.4 GB88 tok/s40K ctx
dense
HelpingAI2 6B i1
C546B6.3 GB72 tok/s101K ctx
dense
Nemotron Mini 4B4B6.3 GB48 tok/s4K ctx
dense
Qwen3.5 9B
C549B8.4 GB88 tok/s40K ctx
dense
Mistral Nemo 12B12B11.7 GB45 tok/s5K ctx
dense
Yi Coder 9B Chat
C549B8.4 GB88 tok/s40K ctx
dense
glm 4 9b chat 1m
C549B8.4 GB88 tok/s40K ctx
dense
Yi 1.5 9B Chat
C549B8.4 GB88 tok/s40K ctx
dense
SOLAR 10.7B Instruct v1.0 uncensored
C5410.7B9.7 GB87 tok/s20K ctx
dense
Nous Hermes 2 SOLAR 10.7B
C5410.7B9.7 GB87 tok/s20K ctx
dense
TinyLlama 1.1B1.1B2.9 GB13 tok/s4K ctx
dense
Yi 9B Coder i1
C539B8.4 GB88 tok/s40K ctx
dense
solar finalised finetuned Model 10.7B i1
C5310.7B9.7 GB87 tok/s20K ctx
dense
HelpingAI2 9B
C539B8.4 GB88 tok/s40K ctx
dense
Helply 10.2b chat i1
C5310.2B9.3 GB91 tok/s25K ctx
dense
SOLAR 10.7B v1.0
C5310.7B9.7 GB87 tok/s20K ctx
dense
HelpingAI2 9B i1
C539B8.4 GB88 tok/s40K ctx
dense
HelpingAI 9B 200k i1
C539B8.4 GB88 tok/s40K ctx
dense
HelpingAI 9B i1
C539B8.4 GB88 tok/s40K ctx
dense
HelpingAI2.5 10B i1
C5310B9.2 GB80 tok/s27K ctx
dense
Gemma 2 2B2B4.7 GB24 tok/s8K ctx
dense
HelpingAI2.5 5B i1
C525B5.5 GB60 tok/s138K ctx
dense
Qwen 2.5 1.5B1.5B3.2 GB18 tok/s131K ctx
dense
Qwen3.5 4B4B4.8 GB48 tok/s193K ctx
dense
gemma 3 4b it
C504B4.8 GB48 tok/s193K ctx
dense
gemma 3 4b it
C504B4.8 GB48 tok/s193K ctx
dense
Gemma 3 1B1B2.9 GB12 tok/s33K ctx
dense
Llama 3.2 3B Instruct
C493B4.4 GB36 tok/s270K ctx
dense
Qwen2.5 3B Instruct3B4.1 GB36 tok/s285K ctx
dense
Llama 3.2 3B Instruct
C483B4.1 GB36 tok/s285K ctx
dense
Ministral 3 3B Instruct 2512
C483B4.1 GB36 tok/s285K ctx
dense
HELVETE 3B
C483B4.1 GB36 tok/s285K ctx
dense
Hermes 3 Llama 3.2 3B
C473B4.1 GB36 tok/s285K ctx
dense
stablelm zephyr 3b
C473B4.1 GB36 tok/s285K ctx
dense
StarCoder2 3B
C473B4.1 GB36 tok/s285K ctx
dense
Qwen 2.5 Coder 0.5B0.5B2.4 GB6 tok/s131K ctx
dense
AI21 Jamba Reasoning 3B
C473B4.1 GB36 tok/s285K ctx
dense
stablelm 3b 4e1t
C473B4.1 GB36 tok/s285K ctx
dense
ai21labs AI21 Jamba Reasoning 3B
C473B4.1 GB36 tok/s285K ctx
dense
ai21labs AI21 Jamba2 3B
C473B4.1 GB36 tok/s285K ctx
dense
HelpingAI 3B hindi i1
C473B4.1 GB36 tok/s285K ctx
dense
AI21 Jamba2 3B
C473B4.1 GB36 tok/s285K ctx
dense
HelpingAI 3B hindi
C473B4.1 GB36 tok/s285K ctx
dense
AI21 Jamba2 3B i1
C473B4.1 GB36 tok/s285K ctx
dense
Qwen 3 0.6B0.6B3.1 GB7 tok/s33K ctx
dense
Llama 3.2 1B1B3.0 GB12 tok/s128K ctx
dense
StarCoder2 3B3B4.2 GB36 tok/s16K ctx
dense
gemma 2 2b it
C462B3.8 GB24 tok/s441K ctx
dense
EXAONE 3.5 2.4B Instruct
C462.4B3.6 GB29 tok/s378K ctx
dense
gemma 2b2B3.4 GB24 tok/s470K ctx
dense
Gemmasutra Mini 2B v1
C452B3.4 GB24 tok/s470K ctx
dense
gemma 2 2b it
C452B3.4 GB24 tok/s470K ctx
dense
Qwen 3.5 0.6B0.6B3.1 GB7 tok/s131K ctx
dense
Qwen2.5 1.5B Instruct1.5B3.0 GB18 tok/s654K ctx
dense
DeepSeek R1 Distill Qwen 1.5B1.5B3.0 GB18 tok/s654K ctx
dense
Yi Coder 1.5B Chat
C431.5B3.0 GB18 tok/s654K ctx
dense
stablelm 2 zephyr 1.6b
C431.6B3.1 GB19 tok/s608K ctx
dense
logos16v2 stablelm2 1.6b i1
C431.6B3.1 GB19 tok/s608K ctx
dense
Yi Coder 1.5B
C431.5B3.0 GB18 tok/s654K ctx
dense
Falcon H1 1.5B Instruct1.5B3.0 GB18 tok/s654K ctx
dense
gemma 3 12b it
C4212B10.6 GB51 tok/s9K ctx
dense
EXAONE 4.0 1.2B
C421.2B2.8 GB14 tok/s838K ctx
dense
TinyLlama 1.1B Chat v1.0
C421.1B2.7 GB13 tok/s922K ctx
dense
Llama 3.2 1B Instruct Q8 0
C421B2.8 GB12 tok/s994K ctx
dense
TinyLlama 1.1B Chat v0.3
C411.1B2.7 GB13 tok/s922K ctx
dense
TinyLlama 1.1B Chat v0.61.1B2.7 GB13 tok/s922K ctx
dense
Qwen 2.5 0.5B0.5B2.4 GB6 tok/s131K ctx
dense
Llama 3.2 1B Instruct
C411B2.6 GB12 tok/s1023K ctx
dense
gemma 3 1b it
C411B2.6 GB12 tok/s1023K ctx
dense
TinyLlama 1.1B Chat v1.0 imatrix
C411.1B2.7 GB13 tok/s922K ctx
dense
Falcon3 1B Instruct abliterated
C411B2.6 GB12 tok/s1023K ctx
dense
F
F Tier -- Can't Run (166 models)
Too heavy for this hardware.
SmolVLM 500M Instruct
D390.5B2.4 GB6 tok/s1230K ctx
dense
embeddinggemma 300M
D370.3B2.2 GB4 tok/s1257K ctx
dense
granite embedding 107m multilingual
D350.11B2.1 GB2 tok/s1286K ctx
dense
Falcon H1 Tiny 90M Instruct
D350.09B2.1 GB2 tok/s1287K ctx
dense
Llama 3.1 70B70B49.5 GB2 tok/s4K ctx
dense
Llama 3.3 70B70B49.5 GB2 tok/s4K ctx
dense
Qwen 2.5 14B14B13.4 GB29 tok/s4K ctx
dense
Qwen 2.5 32B32B25.3 GB3 tok/s4K ctx
dense
Qwen 2.5 72B72B50.7 GB2 tok/s4K ctx
dense
Qwen 2.5 Coder 14B14B13.4 GB29 tok/s4K ctx
dense
Qwen 2.5 Coder 32B32B25.3 GB3 tok/s4K ctx
dense
Qwen 3 14B14B12.9 GB32 tok/s4K ctx
dense
Qwen 3 32B32B25.3 GB3 tok/s4K ctx
dense
Qwen 3 235B A22B235B148.1 GB2 tok/s4K ctx
moe
Qwen 3 30B A3B30.5B22.0 GB9 tok/s4K ctx
moe
Qwen3-Coder 30B A3B Instruct30.5B22.0 GB9 tok/s4K ctx
moe
Qwen3-Coder-Next80B52.2 GB6 tok/s4K ctx
moe
Qwen3-Coder 480B A35B Instruct480B297.6 GB2 tok/s4K ctx
moe
Qwen3-VL 30B A3B Instruct30B21.7 GB14 tok/s4K ctx
moe
Qwen 3.5 27B27B21.5 GB4 tok/s4K ctx
dense
Qwen 3.5 35B A3B35B24.7 GB12 tok/s4K ctx
moe
Qwen 3.6 35B A3B35B27.4 GB11 tok/s4K ctx
moe
Qwen 3.6 27B27B19.3 GB4 tok/s4K ctx
dense
Qwen 3.5 122B A10B122B78.8 GB4 tok/s4K ctx
moe
Qwen 3.5 397B A17B397B246.9 GB2 tok/s4K ctx
moe
Gemma 2 9B9B12.5 GB33 tok/s8K ctx
dense
Gemma 2 27B27B29.6 GB3 tok/s4K ctx
dense
Gemma 3 12B12B14.1 GB22 tok/s4K ctx
dense
Gemma 3 27B27B29.6 GB3 tok/s4K ctx
dense
Ministral 3 14B14B12.9 GB31 tok/s4K ctx
multimodal
Mistral Small 24B24B19.0 GB6 tok/s4K ctx
dense
Mistral Small 3.2 24B24B19.0 GB6 tok/s4K ctx
vision
Mixtral 8x7B47B32.5 GB6 tok/s4K ctx
moe
Mixtral 8x22B141B91.3 GB2 tok/s4K ctx
moe
DeepSeek R1 Distill 14B14B13.4 GB29 tok/s4K ctx
dense
DeepSeek R1 Distill 32B32B25.3 GB3 tok/s4K ctx
dense
Phi-4 14B14B13.5 GB28 tok/s4K ctx
dense
Phi-4-reasoning-plus 14B14.7B13.9 GB25 tok/s4K ctx
dense
Granite 4.1 30B30B24.1 GB3 tok/s4K ctx
dense
OLMo 2 13B13B12.3 GB38 tok/s4K ctx
dense
Command R 35B35B25.7 GB3 tok/s4K ctx
dense
Falcon 40B Instruct40B32.5 GB2 tok/s4K ctx
dense
WizardLM 13B13B22.0 GB11 tok/s4K ctx
dense
StarCoder 7B7B13.5 GB45 tok/s8K ctx
dense
StarCoder 15B15B27.3 GB8 tok/s4K ctx
dense
CodeLlama 7B Instruct7B14.0 GB42 tok/s8K ctx
dense
CodeLlama 13B Instruct13B22.0 GB11 tok/s4K ctx
dense
StarCoder2 15B15B13.9 GB21 tok/s4K ctx
dense
MPT-7B-Instruct7B14.0 GB42 tok/s8K ctx
dense
MPT-30B-Instruct30B46.9 GB3 tok/s4K ctx
dense
InternLM 7B7B14.0 GB42 tok/s8K ctx
dense
InternLM Chat 7B7B14.0 GB42 tok/s8K ctx
dense
Devstral Small 1.124B19.0 GB6 tok/s4K ctx
dense
Devstral Small 2 24B Instruct24B19.0 GB6 tok/s4K ctx
dense
Devstral 2 123B Instruct123B82.3 GB2 tok/s4K ctx
dense
Mistral Small 4 119B119B79.9 GB4 tok/s4K ctx
moe
Leanstral 119B A6B119B83.3 GB4 tok/s4K ctx
moe
Mistral Large 3675B419.0 GB2 tok/s4K ctx
moe
Magistral Small 250724B19.0 GB6 tok/s4K ctx
dense
Codestral 2 25.0822B17.8 GB8 tok/s4K ctx
dense
GLM-5744B474.8 GB2 tok/s4K ctx
moe
GLM-5.1754B480.9 GB2 tok/s4K ctx
moe
Kimi K2.51000B619.3 GB2 tok/s4K ctx
moe
Kimi K2.61000B619.3 GB2 tok/s4K ctx
moe
Kimi Linear 48B A3B48B32.1 GB2 tok/s4K ctx
linear
Nous Hermes 1.09B19.6 GB16 tok/s4K ctx
dense
Nous Dolphin 13B13B23.5 GB9 tok/s4K ctx
dense
Baichuan 7B7B14.0 GB42 tok/s8K ctx
dense
Baichuan 13B13B23.5 GB9 tok/s4K ctx
dense
Cerebras-GPT 13B13B21.0 GB10 tok/s4K ctx
dense
StableLM 2 12B12B22.7 GB9 tok/s4K ctx
dense
InternLM 20B20B34.6 GB5 tok/s4K ctx
dense
Yi 34B Chat34B26.3 GB3 tok/s4K ctx
dense
DeepSeek V3 671B671B470.8 GB2 tok/s4K ctx
moe
DeepSeek V3.1 671B671B470.8 GB2 tok/s4K ctx
moe
DeepSeek V4 Pro1600B865.8 GB2 tok/s4K ctx
moe
DeepSeek V4 Flash284B161.2 GB2 tok/s4K ctx
moe
DeepSeek R1 671B671B470.8 GB2 tok/s4K ctx
moe
DeepSeek Coder V2 16B16B15.0 GB41 tok/s4K ctx
moe
DeepSeek Coder V2 236B236B204.5 GB2 tok/s4K ctx
moe
DeepSeek V2.5 236B236B204.5 GB2 tok/s4K ctx
moe
DeepSeek LLM 7B7B13.5 GB45 tok/s4K ctx
dense
DeepSeek LLM 67B67B48.6 GB2 tok/s4K ctx
dense
Llama 4 Scout 17B 16E109B71.3 GB3 tok/s4K ctx
moe
Llama 4 Maverick 17B 128E400B248.8 GB2 tok/s4K ctx
moe
Command R+ 104B104B68.8 GB2 tok/s4K ctx
dense
Command A 111B111B73.5 GB2 tok/s4K ctx
dense
Phi 3 Medium 14B14B13.5 GB28 tok/s4K ctx
dense
Yi 1.5 34B34B26.3 GB3 tok/s4K ctx
dense
Codestral 22B22B17.8 GB8 tok/s4K ctx
dense
Pixtral Large 124B124B82.9 GB2 tok/s4K ctx
dense
Qwen 2.5 VL 72B72B50.7 GB2 tok/s4K ctx
dense
Qwen 2.5 Math 72B72B50.7 GB2 tok/s4K ctx
dense
Nemotron 70B70B49.5 GB2 tok/s4K ctx
dense
LLaVA 1.5 7B7B14.0 GB42 tok/s4K ctx
dense
LLaVA 1.6 13B13B22.0 GB11 tok/s4K ctx
dense
CogVLM2 19B19B15.9 GB14 tok/s4K ctx
dense
Vicuna 7B7B14.0 GB42 tok/s4K ctx
dense
Vicuna 13B13B22.0 GB11 tok/s4K ctx
dense
Granite Code 20B20B17.3 GB11 tok/s4K ctx
dense
GPT-OSS 20B21B17.2 GB24 tok/s4K ctx
moe
OLMo 2 32B32B25.3 GB3 tok/s4K ctx
dense
Nemotron 3 Nano 30B30B22.6 GB3 tok/s4K ctx
dense
EXAONE 4.0 32B32B25.3 GB3 tok/s4K ctx
dense
LFM2 24B24B19.0 GB6 tok/s4K ctx
dense
Llama 3.1 405B405B256.6 GB2 tok/s4K ctx
dense
DeepSeek R1 Distill 70B70B49.5 GB2 tok/s4K ctx
dense
Granite Code 34B34B26.3 GB3 tok/s4K ctx
dense
GPT-OSS 120B117B78.2 GB2 tok/s4K ctx
dense
Aya Expanse 32B32B23.9 GB3 tok/s4K ctx
dense
Mistral Small 3.1 24B24B19.0 GB6 tok/s4K ctx
dense
Gemma 4 26B A4B25.2B20.9 GB13 tok/s4K ctx
moe
Gemma 4 31B30.7B35.3 GB2 tok/s4K ctx
dense
DeepSeek V3.2671B411.7 GB2 tok/s4K ctx
moe
Nemotron Cascade 2 30B A3B30B23.1 GB13 tok/s4K ctx
moe
MiniMax M2.7
F0230B146.0 GB2 tok/s4K ctx
moe
Solar Open 100B
F0100B74.6 GB2 tok/s4K ctx
dense
Baichuan M2 32B Q4 K M
F032B25.2 GB3 tok/s4K ctx
dense
baichuan inc Baichuan M2 32B
F032B25.2 GB3 tok/s4K ctx
dense
Codestral 22B v0.1
F022B17.9 GB8 tok/s4K ctx
dense
cognitivecomputations Dolphin Mistral 24B Venice Edition
F024B19.4 GB6 tok/s4K ctx
dense
cognitivecomputations Dolphin3.0 R1 Mistral 24B
F024B19.4 GB6 tok/s4K ctx
dense
internlm JanusCoder 14B
F014B12.1 GB33 tok/s4K ctx
dense
internlm2 5 20b chat
F020B16.4 GB11 tok/s4K ctx
dense
NousResearch Hermes 4 14B
F014B12.1 GB33 tok/s4K ctx
dense
starcoder2 15b instruct v0.1
F015B12.8 GB27 tok/s4K ctx
dense
Qwen3 48B A4B Savant Commander Distill 12X Closed Open Heretic Uncensored
F048B36.8 GB2 tok/s4K ctx
dense
internlm2 limarp chat 20b
F020B16.4 GB11 tok/s4K ctx
dense
Phi 4 reasoning vision 15B
F015B12.8 GB27 tok/s4K ctx
dense
Codestral 22B v0.1 IMat
F022B17.9 GB8 tok/s4K ctx
dense
EXAONE 4.0 32B
F032B25.2 GB3 tok/s4K ctx
dense
K EXAONE 236B A23B
F0236B173.5 GB2 tok/s4K ctx
dense
Codestral 22B v0.1
F022B17.9 GB8 tok/s4K ctx
dense
Qwen3.5 35B A3B
F035B27.4 GB3 tok/s4K ctx
dense
starcoder2 15b instruct v0.1
F015B12.8 GB27 tok/s4K ctx
dense
gemma 3 27b it
F027B21.5 GB4 tok/s4K ctx
dense
Llama 3.3 70B Instruct
F070B52.8 GB2 tok/s4K ctx
dense
Mistral Small 24B Instruct 2501
F024B19.4 GB6 tok/s4K ctx
dense
mistral small 3.1 24b instruct 2503 hf
F024B19.4 GB6 tok/s4K ctx
dense
aya expanse 32b heretic MPOA i1
F032B25.2 GB3 tok/s4K ctx
dense
Baichuan M3 235B
F0235B172.8 GB2 tok/s4K ctx
dense
Baichuan M3 235B i1
F0235B172.8 GB2 tok/s4K ctx
dense
BaichuanMed OCR 72B i1
F072B54.3 GB2 tok/s4K ctx
dense
Codestral 21B Pruned i1
F021B17.2 GB9 tok/s4K ctx
dense
Codestral 22B v0.1 i1
F022B17.9 GB8 tok/s4K ctx
dense
Codestral RAG 19B Pruned i1
F019B15.7 GB13 tok/s4K ctx
dense
Dolphin Mistral GLM 4.7 Flash 24B Venice Edition Thinking Uncensored i1
F024B19.4 GB6 tok/s4K ctx
dense
HelpingAI 15B i1
F015B12.8 GB27 tok/s4K ctx
dense
internlm2 math plus 20b i1
F020B16.4 GB11 tok/s4K ctx
dense
Solar Open 100B i1
F0100B74.6 GB2 tok/s4K ctx
dense
Solar Open 69B REAP i1
F069B52.1 GB2 tok/s4K ctx
dense
starcoder2 15b i1
F015B12.8 GB27 tok/s4K ctx
dense
Hermes 4.3 36B
F036B28.1 GB2 tok/s4K ctx
dense
stabilityai japanese stablelm instruct beta 70b
F070B52.8 GB2 tok/s4K ctx
dense
Codestral 22B v0.1
F022B17.9 GB8 tok/s4K ctx
dense
StarCoder2 15B
F015B12.8 GB27 tok/s4K ctx
dense
GGUF SOLARized GraniStral 14B 1902 YeAM HCT
F014B12.1 GB33 tok/s4K ctx
dense
GGUF SOLARized GraniStral 14B 2102 YeAM HCT 32QKV
F014B12.1 GB33 tok/s4K ctx
dense
DeepSeek R1 Distill Qwen 14B14B12.1 GB33 tok/s4K ctx
dense
gemma 3 27b it27B21.5 GB4 tok/s4K ctx
dense
Mistral Small 3.2 24B Instruct 250624B19.4 GB6 tok/s4K ctx
dense
Qwen3.5 122B A10B122B76.0 GB2 tok/s4K ctx
dense
Qwen3.5 27B27B21.5 GB4 tok/s4K ctx
dense
Qwen3.5 35B A3B35B27.4 GB3 tok/s4K ctx
dense
Qwen3.5 397B A17B397B290.6 GB2 tok/s4K ctx
dense
cognitivecomputations Dolphin Mistral 24B Venice Edition
F024B19.4 GB6 tok/s4K ctx
dense
Compare with...
See How Other Hardware Stacks Up
Check tier lists for similar GPUs and Apple Silicon configs