Index/Leaderboard
Index / Open-weight models on Apple Silicon / 18 November 2026

The local LLM leaderboard

According to the LLMCheck index (18 November 2026), MiMo-V2.6-Flash and Qwen 3.6-35B-A3B tie for #1 of 130 open-weight models at Score 81 (MiMo-V2.6-Flash is listed first on the alphabetical tie-break). MiMo-V2.6-Flash needs a Mac Studio with 192 GB or more; Qwen 3.6-35B-A3B needs 24 GB, the best pick at 32 GB or less. Every speed carries a provenance mark.

Provenance key
sourced
Published figure — vendor or press review
estimated
community
Linked individual run — blog, forum or submission
Fits in License 130 of 130 shown
Use
# Model Score↓0–100 composite Speedtok/s decode RAM @64Kagent context RAMunified min License
1 MiMo-V2.6-Flash
309B MoE · MXFP4 · Xiaomi
81
community77 M5 Ultracommunityprefill 2,755 @8KcommunityTTFT 0.609 s @1K estimated172 GB 170 GB MIT
2 Qwen 3.6-35B-A3B
35B MoE · 4-bit · Alibaba
ollama run qwen3.6:35b-a3bHugging Face
81
community130communityprefill 4,576 @1KcommunityTTFT 0.224 s @1K estimated26 GB 24 GB Apache 2.0
3 GLM-4.7-Flash
31B MoE · Q4_K_M · Zhipu AI
ollama run glm-4.7-flashHugging Face
80
estimated107 estimated22 GB 18 GB MIT
4 Recursive Flint
39.8B MoE · 4-bit · Recursive
ollama run flint:40bHugging Face
80
community123communityprefill 3,592 @8KcommunityTTFT 0.309 s @1K estimated22 GB 20 GB Apache 2.0
5 KAT-Coder-V2.5
35B MoE · OptiQ-4bit · Kwaipilot
79
community89communityprefill 4,018 @4KcommunityTTFT 0.316 s @1K estimated26 GB 24 GB Apache 2.0
6 Nemotron 3.5 Lightning
30B MoE · Q4_0 · NVIDIA
ollama run nemotron-3.5-lightning:30bHugging Face
79
community117 estimated19 GB 18 GB OpenMDW
7 North Mini Code 1.0
30.5B MoE · 4-bit · Cohere
ollama run north-mini-code-1.0Hugging Face
79
community120communityprefill 3,424 @8KcommunityTTFT 0.376 s @1K estimated23 GB 21 GB Apache 2.0
8 Gemma 4 26B-A4B
26B MoE · Q4_K_M · Google
ollama run gemma4:26b-a4bHugging Face
78
sourced107communityprefill 3,852 @4K estimated20 GB 18 GB Apache 2.0
9 MiniCPM-V 4.7 35B-A3B
35.2B MoE · MXFP4 · OpenBMB
ollama run openbmb/minicpm-v4.7:35b-a3bHugging Face
78
community100communityprefill 3,583 @4KcommunityTTFT 0.804 s @1K estimated25 GB 23 GB Apache 2.0
10 Raon-1 mini preview
28.8B MoE · Q4_K_M · KRAFTON
77
estimated123 estimated20 GB 18 GB Apache 2.0
11 LFM2.5-24B-A2B
24B MoE · Q4_K_M · Liquid AI
ollama run lfm2.5:24bHugging Face
76
sourced184communityprefill 3,772 @4K estimated19 GB 17 GB LFM Open
12 Laguna XS 2.1
33B MoE · Q4_K_M · Poolside
76
estimated117 estimated27 GB 24 GB OpenMDW
13 Bonsai 27B
27B · Q1_0_g128 · Prism ML
72
sourced66 estimated11 GB 6 GB Apache 2.0
14 Qwen3.8-27B
27.8B · 4-bit · Alibaba
ollama run qwen3.8:27bHugging Face
72
community31communityprefill 839 @4K estimated24 GB 19 GB Apache 2.0
15 Nemotron-Cascade 2
30B MoE · Q4_K_M · NVIDIA
ollama run nemotron-cascade-2:30bHugging Face
71
estimated117 estimated19 GB 18 GB NVIDIA Open
16 DeepSeek V4 Flash
284B MoE · Q2_K · DeepSeek
70
community39communityTTFT 0.4 s estimated121 GB 120 GB MIT
17 DeepSeek V4 Flash Vision-Exp
305B MoE · Q2_K · DeepSeek
70
estimated39 estimated121 GB 120 GB MIT
18 DeepSeek V4.1 Flash
552B MoE · oQ4e · DeepSeek
70
estimated29 M5 Ultra estimated435 GB 434 GB MIT
19 K2-Horizon-3.7B
5.1B · Q4_K_M · MBZUAI IFM
70
estimated133 estimated14 GB 4 GB Apache 2.0
20 Ling-3.0-flash-VL
124B MoE · Q4_K_M · Ant Group
70
estimated53 estimated57 GB 56 GB MIT
21 Maple Preview 20B-A1B
20B MoE · Ternary · DeepGrove
70
sourced387 M5 Pro estimated7 GB 6 GB MIT
22 Nanbeige4.3-33B
33.1B · oQ4e · Nanbeige
70
community25communityprefill 613 @4KcommunityTTFT 1.665 s @1K estimated31 GB 17 GB Apache 2.0
23 Nemotron 3.5 Bolt 74B-A9B
74B MoE · 5-bit · NVIDIA
70
community64communityprefill 1,332 @4K estimated46 GB 45 GB OpenMDW
24 Qwen 3.5 35B-A3B
35B MoE · Q4_K_M · Alibaba
ollama run qwen3.5:35b-a3bHugging Face
70
estimated136 estimated26 GB 24 GB Apache 2.0
25 Qwen 3.6-27B
27B · Q4_K_M · Alibaba
ollama run qwen3.6:27bHugging Face
70
estimated30 estimated23 GB 18 GB Apache 2.0
26 Qwen3-Coder-Next
80B MoE · Q4_K_M · Alibaba
ollama run qwen3-coder-nextHugging Face
70
estimated84 estimated50 GB 48 GB Apache 2.0
27 GLM 5.2
753B MoE · Q4_K_M · Zhipu AI
69
estimated22 M5 Ultra estimated262 GB 256 GB MIT
28 MiniCPM5-2B
2.52B · Q4_K_M · OpenBMB
69
estimated225 estimated6 GB 3 GB Apache 2.0
29 Ternary Bonsai 2 27B
27.36B · PQ2_0 · Prism ML
ollama run bonsai-2:27bHugging Face
69
sourced47 estimated14 GB 9 GB Apache 2.0
30 GLM-5.3-Flash
320B MoE · IQ3_XXS · Zhipu AI
68
estimated27 estimated129 GB 128 GB MIT
31 GPT-oss 120B
117B MoE · MXFP4 · OpenAI
ollama run gpt-oss:120bHugging Face
68
community92communityprefill 2,022 @4KcommunityTTFT 1.109 s @1K estimated65 GB 62 GB Apache 2.0
32 Inkling 2 Small
594B MoE · Q4_K_M · Thinking Machines
68
estimated36 M5 Ultra estimated231 GB 228 GB Apache 2.0
33 Kolibri-1
78.1B MoE · Q4_K_M · Aleph Alpha
68
estimated75 estimated52 GB 50 GB Apache 2.0
34 Ling-3.0 Flash
124B MoE · 4-bit · Ant Group
68
community68communityprefill 1,615 @4KcommunityTTFT 0.644 s @1K estimated97 GB 96 GB MIT
35 Nanbeige4.2-3B
3B · Q4_K_M · Nanbeige
68
estimated199 estimated8 GB 2 GB Apache 2.0
36 Spark-X2.6-4B
4.11B · oQ5e · SparkLLM (XHToken)
68
community102communityprefill 2,867 @8KcommunityTTFT 0.237 s @1K estimated7 GB 4 GB Apache 2.0
37 GPT-oss 20B
21B MoE · MXFP4 · OpenAI
ollama run gpt-oss:20bHugging Face
67
community141communityprefill 4,774 @4KcommunityTTFT 0.587 s @1K estimated15 GB 13 GB Apache 2.0
38 Inkling-Small
276B MoE · Q2_K · Thinking Machines
67
estimated42 estimated98 GB 96 GB Apache 2.0
39 Kimi K2.6
1T MoE · 3.6 bpw · Moonshot AI
67
estimated27 M5 Ultra estimated434 GB 429 GB Modified MIT
40 Kimi K2.7 Code
1T MoE · IQ3_S · Moonshot AI
67
estimated27 M5 Ultra estimated426 GB 421 GB Modified MIT
41 Ling-3.1-flash
560B MoE · Q4_K_M · Ant Group
67
estimated27 M5 Ultra estimated361 GB 360 GB MIT
42 Muse Glimmer 30B
30B · Q4_K_M · Meta
67
sourced27sourcedTTFT 0.4 s estimated19 GB 18 GB Apache 2.0
43 Qwen3.8-Flash-Next
125B MoE · 4-bit · Alibaba
67
community47communityprefill 1,309 @1K estimated122 GB 120 GB Qwen Custom
44 GLM-5.3
753B MoE · Q4_K_M · Zhipu AI
66
estimated17 M5 Ultra estimated262 GB 256 GB GLM-5.3
45 Gemma 4 12B
12B · Q4_K_M · Google
ollama run gemma4:12bHugging Face
66
estimated64 estimated11 GB 9 GB Apache 2.0
46 Kimi K2.5
1T MoE · 3.5 bpw · Moonshot AI
66
estimated28 M5 Ultra estimated422 GB 417 GB Modified MIT
47 MiMo-V2.6-Distill-Qwen-9B
9.41B · Q4_K_M · Xiaomi
ollama run mimo-v2.6:9bHugging Face
66
estimated79 estimated11 GB 8 GB MIT
48 MiniMax M2.5
230B MoE · Q4_K_M · MiniMax
66
estimated57 M5 Ultra estimated146 GB 130 GB Modified MIT
49 Beam 1.0
501.3B MoE · Q4_K_M · Reflection AI
65
estimated27 M5 Ultra estimated202 GB 199 GB Apache 2.0
50 GLM-5.1
754B MoE · Q4_K_M · Zhipu AI
65
estimated17 M5 Ultra estimated396 GB 390 GB MIT
51 Ling-3.0-tiny
7.9B MoE · oQ4e · Ant Group
65
community174communityprefill 2,012 @4KcommunityTTFT 2.035 s @4K estimated7 GB 6 GB MIT
52 MiniMax-M3
428B MoE · Q4_K_M · MiniMax
65
estimated42 M5 Ultra estimated170 GB 162 GB MiniMax Community
53 Muse Spark Open
816B MoE · oQ3e · Meta
65
estimated31 M5 Ultra — 342 GB Muse Model License
54 Nemotron 3.5 Bolt E4B
4B · Q4_K_M · NVIDIA
ollama run nemotron-3.5-bolt:e4bHugging Face
65
estimated161 estimated8 GB 6 GB OpenMDW
55 Qwen 3 30B-A3B
30B MoE · 4-bit · Alibaba
ollama run qwen3:30b-a3bHugging Face
65
sourced152 estimated25 GB 18 GB Apache 2.0
56 Qwen 3.5 397B-A17B
397B MoE · 4-bit · Alibaba
65
community62 M5 Ultracommunityprefill 1,265 @8K estimated212 GB 210 GB Apache 2.0
57 DeepSeek R1 8B
8B · Q4_K_M · DeepSeek
ollama run deepseek-r1:8bHugging Face
64
estimated91 estimated16 GB 6 GB MIT
58 Gemma 4 31B
31B · 4-bit · Google
ollama run gemma4:31bHugging Face
64
community26communityprefill 710 @4KcommunityTTFT 1.439 s @1K estimated30 GB 24 GB Apache 2.0
59 Gemma 4 E4B
4B · Q4_K_M · Google
ollama run gemma4:e4bHugging Face
64
estimated161 estimated8 GB 6 GB Apache 2.0
60 Laguna S 2.1
118B MoE · Q4_K_M · Poolside
64
estimated38 estimated100 GB 96 GB OpenMDW
61 MiniMax-M3.1-Flash
428B MoE · oQ3e · MiniMax
64
community35 M5 Ultracommunityprefill 1,110 @4KcommunityTTFT 0.969 s @1K estimated197 GB 189 GB MiniMax Community
62 Phi-4 Mini
3.8B · Q4_K_M · Microsoft
ollama run phi4-mini:3.8bHugging Face
64
estimated168 estimated12 GB 3 GB MIT
63 Step-3.7-Flash
198B MoE · oQ2e · StepFun
64
community48communityprefill 594 @4KcommunityTTFT 1.704 s @1K estimated69 GB 65 GB Apache 2.0
64 DeepSeek V4 Pro
1.6T MoE · DeepSeek
63
— estimated851 GB 850 GB MIT
65 Hy4 preview
770B MoE · Q4_K_M · Tencent
63
estimated14 M5 Ultra estimated298 GB 292 GB Apache 2.0
66 LFM3-1.7B
1.73B · Q4_K_M · Liquid AI
ollama run lfm3:1.7bHugging Face
63
estimated285 estimated4 GB 3 GB LFM Open
67 Mellum2.1-12B-A2.5B-Thinking
12.15B MoE · Q4_K_M · JetBrains
63
estimated154 estimated11 GB 10 GB Apache 2.0
68 MiMo-V2.6-Pro
1.02T MoE · Xiaomi
63
— estimated585 GB 581 GB MIT
69 Qwen3-235B-A22B
235B MoE · 4.5 bpw · Alibaba
ollama run qwen3:235b-a22bHugging Face
63
estimated26 estimated140 GB 128 GB Apache 2.0
70 Apertus 1.5 8B
8B · Q4_K_M · Swiss AI
62
estimated91 — 5 GB Apache 2.0
71 DeepSeek V3.2-Speciale
685B MoE · Q4_K_M · DeepSeek
62
estimated18 M5 Ultra estimated365 GB 360 GB MIT
72 Nemotron 3.5 Super VL 120B-A12B
120B MoE · Q4_0 · NVIDIA
ollama run nemotron-3.5-super:120bHugging Face
62
community36communityprefill 885 @4K estimated73 GB 72 GB OpenMDW
73 dots3-note preview
280B MoE · 3.5 bpw · Xiaohongshu (RedNote)
62
estimated28 estimated121 GB 120 GB Apache 2.0
74 Gemma 4 E2B
2B · Q4_K_M · Google
ollama run gemma4:e2bHugging Face
61
estimated261 estimated3 GB 2 GB Apache 2.0
75 Mistral 7B
7B · Q4_K_M · Mistral
ollama run mistral:7bHugging Face
61
estimated102 32K max 5 GB Apache 2.0
76 Qwen 3 8B
8B · Q4_K_M · Alibaba
ollama run qwen3:8bHugging Face
61
estimated91 estimated16 GB 6 GB Apache 2.0
77 Qwen 3.5 9B
9B · Q4_K_M · Alibaba
ollama run qwen3.5:9bHugging Face
61
estimated81 estimated10 GB 7 GB Apache 2.0
78 Granite 4.2 8B
8.79B · Q4_K_M · IBM
ollama run granite4.2:8bHugging Face
60
estimated84 estimated19 GB 8 GB Apache 2.0
79 Hunyuan Hy3
295B MoE · Q4_K_M · Tencent
60
estimated30 M5 Ultra estimated181 GB 160 GB Apache 2.0
80 LFM2.5-2.6B
2.6B · Q4_K_M · Liquid AI
60
sourced220sourcedTTFT 0.1 s estimated4 GB 2 GB LFM Open
81 Ministral 8B
8B · Q4_K_M · Mistral
60
estimated91 estimated12 GB 6 GB Apache 2.0
82 Qwen 3.5 122B-A10B
122B MoE · Q4_K_M · Alibaba
ollama run qwen3.5:122b-a10bHugging Face
60
community43 estimated67 GB 65 GB Apache 2.0
83 Qwen 3.5 4B
4B · Q4_K_M · Alibaba
ollama run qwen3.5:4bHugging Face
60
estimated161 estimated6 GB 3 GB Apache 2.0
84 DeepSeek V3.2
685B MoE · Q4_K_M · DeepSeek
59
estimated18 M5 Ultra estimated365 GB 360 GB MIT
85 Mistral Small 4
119B MoE · Q4_K_M · Mistral
59
estimated46 estimated66 GB 64 GB Apache 2.0
86 Qwen 3 4B
4B · Q4_K_M · Alibaba
ollama run qwen3:4bHugging Face
59
estimated161 estimated13 GB 3 GB Apache 2.0
87 Ripple-65B Preview
64.8B · Q4_K_M · Ricursive Intelligence
ollama run ripple-preview:65bHugging Face
59
estimated13 estimated38 GB 33 GB Apache 2.0
88 Inkling
975B MoE · Thinking Machines
58
— estimated553 GB 550 GB Apache 2.0
89 Kimi K3
2.8T MoE · Moonshot AI
58
— estimated1562 GB 1560 GB Moonshot Custom
90 Kimi K3.1 Pro
2.8T MoE · Moonshot AI
58
— estimated1565 GB 1563 GB Moonshot Custom
91 MiniMax-M2.7
228.7B MoE · IQ3_XXS · MiniMax
58
estimated45 estimated98 GB 82 GB MiniMax Non-Commercial
92 Mistral Large 4
1.05T MoE · 3.0 bpw · Mistral
58
estimated21 M5 Ultra estimated388 GB 383 GB Mistral Ecosystem Open-Weight
93 Qwen3.8-2.4T-A95B
2.4T MoE · Alibaba
58
— estimated1316 GB 1310 GB Qwen Custom
94 SmolLM3 3B
3B · Q4_K_M · HuggingFace
58
estimated199 estimated7 GB 2 GB Apache 2.0
95 Ternary Bonsai 2 9B
9.41B · PQ2_0 · Prism ML
ollama run bonsai-2:9bHugging Face
58
sourced77 M5 Pro estimated8 GB 5 GB Apache 2.0
96 Llama 3.1 8B
8B · 4-bit · Meta
ollama run llama3.1:8bHugging Face
57
sourced102 — 5 GB Meta Custom
97 Step 5 Preview
600B MoE · StepFun
57
— estimated266 GB 261 GB StepFun Model License
98 Gemma 3 4B
4B · Q4_K_M · Google
ollama run gemma3:4bHugging Face
56
estimated161 — 3 GB Gemma
99 Granite 4.2 30B
29.28B · Q4_K_M · IBM
ollama run granite4.2:30bHugging Face
56
estimated28 estimated37 GB 20 GB Apache 2.0
100 Nemotron 3 Ultra 550B-A55B
550B MoE · Q4_K_M · NVIDIA
56
estimated12 M5 Ultra estimated227 GB 226 GB OpenMDW
101 GLM-4.5-Air
106B MoE · Q4_K_M · Zhipu AI
55
estimated27 estimated68 GB 56 GB MIT
102 MiMo-V2-Flash
309B MoE · Q4_K_M · Xiaomi
55
estimated41 M5 Ultra estimated167 GB 165 GB MIT
103 Phi-4 14B
14B · Q4_K_M · Microsoft
ollama run phi4:14bHugging Face
55
estimated55 estimated23 GB 10 GB MIT
104 DeepSeek R1
671B MoE · Q4_K_M · DeepSeek
ollama run deepseek-r1:671bHugging Face
54
estimated18 M5 Ultra estimated355 GB 350 GB MIT
105 DeepSeek V3
685B MoE · Q4_K_M · DeepSeek
ollama run deepseek-v3:671bHugging Face
54
estimated18 M5 Ultra estimated365 GB 360 GB MIT
106 Qwen 3 14B
14B · Q4_K_M · Alibaba
ollama run qwen3:14bHugging Face
54
estimated55 estimated21 GB 10 GB Apache 2.0
107 Llama 4 Maverick
400B MoE · Q4_K_M · Meta
ollama run llama4:maverickHugging Face
53
estimated36 M5 Ultra — 210 GB Meta Custom
108 Mistral Medium 3.5
128B · Q4_K_M · Mistral
ollama run mistral-medium-3.5:128bHugging Face
53
estimated7 estimated100 GB 77 GB Mistral Modified MIT
109 Mistral Small 3.2 24B
24B · Q4_K_M · Mistral
ollama run mistral-small3.2:24bHugging Face
53
estimated33 estimated25 GB 14 GB Apache 2.0
110 Devstral Small 24B
24B · Q4_K_M · Mistral
52
estimated33 estimated25 GB 14 GB Apache 2.0
111 GLM-4.7
355B MoE · Q4_K_M · Zhipu AI
52
estimated21 M5 Ultra estimated204 GB 180 GB MIT
112 Ministral 3 14B
14B · Q4_K_M · Mistral
ollama run ministral-3:14bHugging Face
52
estimated55 estimated21 GB 10 GB Apache 2.0
113 Qwen 2.5 14B
14B · Q4_K_M · Alibaba
ollama run qwen2.5:14bHugging Face
52
estimated55 estimated23 GB 10 GB Apache 2.0
114 Gemma 3 12B
12B · Q4_K_M · Google
ollama run gemma3:12bHugging Face
51
estimated64 — 9 GB Gemma
115 Mistral Large 3
675B MoE · Q4_K_M · Mistral
51
estimated16 M5 Ultra estimated360 GB 355 GB Apache 2.0
116 Solar Open 2 250B
250B MoE · Upstage
51
— estimated164 GB 160 GB Solar License
117 Step-3.5-Flash
196B MoE · Q4_K_M · StepFun
51
estimated29 estimated124 GB 120 GB Apache 2.0
118 DeepSeek R1 32B
32B · Q4_K_M · DeepSeek
ollama run deepseek-r1:32bHugging Face
50
estimated25 estimated39 GB 22 GB MIT
119 QwQ 32B
32B · Q4_K_M · Alibaba
ollama run qwq:32bHugging Face
50
estimated25 32K max 22 GB Apache 2.0
120 DeepSeek R1 70B
70B · Q4_K_M · DeepSeek
ollama run deepseek-r1:70bHugging Face
49
estimated12 estimated65 GB 44 GB MIT
121 Qwen 3 32B
32B · 4-bit · Alibaba
ollama run qwen3:32bHugging Face
49
sourced26 estimated39 GB 22 GB Apache 2.0
122 Qwen 3.5 27B
27B · Q4_K_M · Alibaba
ollama run qwen3.5:27bHugging Face
47
estimated30 estimated23 GB 18 GB Apache 2.0
123 Apertus 1.5 70B
70B · Q4_K_M · Swiss AI
46
estimated12 — 44 GB Apache 2.0
124 Gemma 3 27B
27B · Q4_K_M · Google
ollama run gemma3:27bHugging Face
46
estimated30 — 18 GB Gemma
125 Llama 4 Scout
109B MoE · Q4_K_M · Meta
ollama run llama4:scoutHugging Face
45
estimated19 — 58 GB Meta Custom
126 Qwen 2.5 72B
72B · Q4_K_M · Alibaba
ollama run qwen2.5:72bHugging Face
45
estimated12 estimated67 GB 46 GB Apache 2.0
127 Llama 3.1 405B
405B · Q4_K_M · Meta
ollama run llama3.1:405bHugging Face
44
estimated4 M5 Ultra — 240 GB Meta Custom
128 Hermes 4 70B
70B · Q4_K_M · Nous Research
42
estimated12 estimated59 GB 38 GB Meta Custom
129 Llama 3.3 70B
70B · 4-bit · Meta
ollama run llama3.3:70bHugging Face
41
sourced13sourcedprefill 306 @0.5K — 44 GB Meta Custom
130 Mixtral 8x22B
141B MoE · Q4_K_M · Mistral
ollama run mixtral:8x22bHugging Face
41
estimated9 estimated103 GB 88 GB Apache 2.0
81
#1
MiMo-V2.6-Flash
309B MoE · MXFP4 · Xiaomi · MIT
tok/s
community77 M5 Ultra
RAM @64K
estimated172 GB
Min RAM
170 GB
communityprefill 2,755 @8KcommunityTTFT 0.609 s @1K
81
#2
Qwen 3.6-35B-A3B
35B MoE · 4-bit · Alibaba · Apache 2.0
tok/s
community130
RAM @64K
estimated26 GB
Min RAM
24 GB
communityprefill 4,576 @1KcommunityTTFT 0.224 s @1K
ollama run qwen3.6:35b-a3bHugging Face
80
#3
GLM-4.7-Flash
31B MoE · Q4_K_M · Zhipu AI · MIT
tok/s
estimated107
RAM @64K
estimated22 GB
Min RAM
18 GB
ollama run glm-4.7-flashHugging Face
80
#4
Recursive Flint
39.8B MoE · 4-bit · Recursive · Apache 2.0
tok/s
community123
RAM @64K
estimated22 GB
Min RAM
20 GB
communityprefill 3,592 @8KcommunityTTFT 0.309 s @1K
ollama run flint:40bHugging Face
79
#5
KAT-Coder-V2.5
35B MoE · OptiQ-4bit · Kwaipilot · Apache 2.0
tok/s
community89
RAM @64K
estimated26 GB
Min RAM
24 GB
communityprefill 4,018 @4KcommunityTTFT 0.316 s @1K
79
#6
Nemotron 3.5 Lightning
30B MoE · Q4_0 · NVIDIA · OpenMDW
tok/s
community117
RAM @64K
estimated19 GB
Min RAM
18 GB
ollama run nemotron-3.5-lightning:30bHugging Face
79
#7
North Mini Code 1.0
30.5B MoE · 4-bit · Cohere · Apache 2.0
tok/s
community120
RAM @64K
estimated23 GB
Min RAM
21 GB
communityprefill 3,424 @8KcommunityTTFT 0.376 s @1K
ollama run north-mini-code-1.0Hugging Face
78
#8
Gemma 4 26B-A4B
26B MoE · Q4_K_M · Google · Apache 2.0
tok/s
sourced107
RAM @64K
estimated20 GB
Min RAM
18 GB
communityprefill 3,852 @4K
ollama run gemma4:26b-a4bHugging Face
78
#9
MiniCPM-V 4.7 35B-A3B
35.2B MoE · MXFP4 · OpenBMB · Apache 2.0
tok/s
community100
RAM @64K
estimated25 GB
Min RAM
23 GB
communityprefill 3,583 @4KcommunityTTFT 0.804 s @1K
ollama run openbmb/minicpm-v4.7:35b-a3bHugging Face
77
#10
Raon-1 mini preview
28.8B MoE · Q4_K_M · KRAFTON · Apache 2.0
tok/s
estimated123
RAM @64K
estimated20 GB
Min RAM
18 GB
76
#11
LFM2.5-24B-A2B
24B MoE · Q4_K_M · Liquid AI · LFM Open
tok/s
sourced184
RAM @64K
estimated19 GB
Min RAM
17 GB
communityprefill 3,772 @4K
ollama run lfm2.5:24bHugging Face
76
#12
Laguna XS 2.1
33B MoE · Q4_K_M · Poolside · OpenMDW
tok/s
estimated117
RAM @64K
estimated27 GB
Min RAM
24 GB
72
#13
Bonsai 27B
27B · Q1_0_g128 · Prism ML · Apache 2.0
tok/s
sourced66
RAM @64K
estimated11 GB
Min RAM
6 GB
72
#14
Qwen3.8-27B
27.8B · 4-bit · Alibaba · Apache 2.0
tok/s
community31
RAM @64K
estimated24 GB
Min RAM
19 GB
communityprefill 839 @4K
ollama run qwen3.8:27bHugging Face
71
#15
Nemotron-Cascade 2
30B MoE · Q4_K_M · NVIDIA · NVIDIA Open
tok/s
estimated117
RAM @64K
estimated19 GB
Min RAM
18 GB
ollama run nemotron-cascade-2:30bHugging Face
70
#16
DeepSeek V4 Flash
284B MoE · Q2_K · DeepSeek · MIT
tok/s
community39
RAM @64K
estimated121 GB
Min RAM
120 GB
communityTTFT 0.4 s
70
#17
DeepSeek V4 Flash Vision-Exp
305B MoE · Q2_K · DeepSeek · MIT
tok/s
estimated39
RAM @64K
estimated121 GB
Min RAM
120 GB
70
#18
DeepSeek V4.1 Flash
552B MoE · oQ4e · DeepSeek · MIT
tok/s
estimated29 M5 Ultra
RAM @64K
estimated435 GB
Min RAM
434 GB
70
#19
K2-Horizon-3.7B
5.1B · Q4_K_M · MBZUAI IFM · Apache 2.0
tok/s
estimated133
RAM @64K
estimated14 GB
Min RAM
4 GB
70
#20
Ling-3.0-flash-VL
124B MoE · Q4_K_M · Ant Group · MIT
tok/s
estimated53
RAM @64K
estimated57 GB
Min RAM
56 GB
70
#21
Maple Preview 20B-A1B
20B MoE · Ternary · DeepGrove · MIT
tok/s
sourced387 M5 Pro
RAM @64K
estimated7 GB
Min RAM
6 GB
70
#22
Nanbeige4.3-33B
33.1B · oQ4e · Nanbeige · Apache 2.0
tok/s
community25
RAM @64K
estimated31 GB
Min RAM
17 GB
communityprefill 613 @4KcommunityTTFT 1.665 s @1K
70
#23
Nemotron 3.5 Bolt 74B-A9B
74B MoE · 5-bit · NVIDIA · OpenMDW
tok/s
community64
RAM @64K
estimated46 GB
Min RAM
45 GB
communityprefill 1,332 @4K
70
#24
Qwen 3.5 35B-A3B
35B MoE · Q4_K_M · Alibaba · Apache 2.0
tok/s
estimated136
RAM @64K
estimated26 GB
Min RAM
24 GB
ollama run qwen3.5:35b-a3bHugging Face
70
#25
Qwen 3.6-27B
27B · Q4_K_M · Alibaba · Apache 2.0
tok/s
estimated30
RAM @64K
estimated23 GB
Min RAM
18 GB
ollama run qwen3.6:27bHugging Face
70
#26
Qwen3-Coder-Next
80B MoE · Q4_K_M · Alibaba · Apache 2.0
tok/s
estimated84
RAM @64K
estimated50 GB
Min RAM
48 GB
ollama run qwen3-coder-nextHugging Face
69
#27
GLM 5.2
753B MoE · Q4_K_M · Zhipu AI · MIT
tok/s
estimated22 M5 Ultra
RAM @64K
estimated262 GB
Min RAM
256 GB
69
#28
MiniCPM5-2B
2.52B · Q4_K_M · OpenBMB · Apache 2.0
tok/s
estimated225
RAM @64K
estimated6 GB
Min RAM
3 GB
69
#29
Ternary Bonsai 2 27B
27.36B · PQ2_0 · Prism ML · Apache 2.0
tok/s
sourced47
RAM @64K
estimated14 GB
Min RAM
9 GB
ollama run bonsai-2:27bHugging Face
68
#30
GLM-5.3-Flash
320B MoE · IQ3_XXS · Zhipu AI · MIT
tok/s
estimated27
RAM @64K
estimated129 GB
Min RAM
128 GB
68
#31
GPT-oss 120B
117B MoE · MXFP4 · OpenAI · Apache 2.0
tok/s
community92
RAM @64K
estimated65 GB
Min RAM
62 GB
communityprefill 2,022 @4KcommunityTTFT 1.109 s @1K
ollama run gpt-oss:120bHugging Face
68
#32
Inkling 2 Small
594B MoE · Q4_K_M · Thinking Machines · Apache 2.0
tok/s
estimated36 M5 Ultra
RAM @64K
estimated231 GB
Min RAM
228 GB
68
#33
Kolibri-1
78.1B MoE · Q4_K_M · Aleph Alpha · Apache 2.0
tok/s
estimated75
RAM @64K
estimated52 GB
Min RAM
50 GB
68
#34
Ling-3.0 Flash
124B MoE · 4-bit · Ant Group · MIT
tok/s
community68
RAM @64K
estimated97 GB
Min RAM
96 GB
communityprefill 1,615 @4KcommunityTTFT 0.644 s @1K
68
#35
Nanbeige4.2-3B
3B · Q4_K_M · Nanbeige · Apache 2.0
tok/s
estimated199
RAM @64K
estimated8 GB
Min RAM
2 GB
68
#36
Spark-X2.6-4B
4.11B · oQ5e · SparkLLM (XHToken) · Apache 2.0
tok/s
community102
RAM @64K
estimated7 GB
Min RAM
4 GB
communityprefill 2,867 @8KcommunityTTFT 0.237 s @1K
67
#37
GPT-oss 20B
21B MoE · MXFP4 · OpenAI · Apache 2.0
tok/s
community141
RAM @64K
estimated15 GB
Min RAM
13 GB
communityprefill 4,774 @4KcommunityTTFT 0.587 s @1K
ollama run gpt-oss:20bHugging Face
67
#38
Inkling-Small
276B MoE · Q2_K · Thinking Machines · Apache 2.0
tok/s
estimated42
RAM @64K
estimated98 GB
Min RAM
96 GB
67
#39
Kimi K2.6
1T MoE · 3.6 bpw · Moonshot AI · Modified MIT
tok/s
estimated27 M5 Ultra
RAM @64K
estimated434 GB
Min RAM
429 GB
67
#40
Kimi K2.7 Code
1T MoE · IQ3_S · Moonshot AI · Modified MIT
tok/s
estimated27 M5 Ultra
RAM @64K
estimated426 GB
Min RAM
421 GB
67
#41
Ling-3.1-flash
560B MoE · Q4_K_M · Ant Group · MIT
tok/s
estimated27 M5 Ultra
RAM @64K
estimated361 GB
Min RAM
360 GB
67
#42
Muse Glimmer 30B
30B · Q4_K_M · Meta · Apache 2.0
tok/s
sourced27
RAM @64K
estimated19 GB
Min RAM
18 GB
sourcedTTFT 0.4 s
67
#43
Qwen3.8-Flash-Next
125B MoE · 4-bit · Alibaba · Qwen Custom
tok/s
community47
RAM @64K
estimated122 GB
Min RAM
120 GB
communityprefill 1,309 @1K
66
#44
GLM-5.3
753B MoE · Q4_K_M · Zhipu AI · GLM-5.3
tok/s
estimated17 M5 Ultra
RAM @64K
estimated262 GB
Min RAM
256 GB
66
#45
Gemma 4 12B
12B · Q4_K_M · Google · Apache 2.0
tok/s
estimated64
RAM @64K
estimated11 GB
Min RAM
9 GB
ollama run gemma4:12bHugging Face
66
#46
Kimi K2.5
1T MoE · 3.5 bpw · Moonshot AI · Modified MIT
tok/s
estimated28 M5 Ultra
RAM @64K
estimated422 GB
Min RAM
417 GB
66
#47
MiMo-V2.6-Distill-Qwen-9B
9.41B · Q4_K_M · Xiaomi · MIT
tok/s
estimated79
RAM @64K
estimated11 GB
Min RAM
8 GB
ollama run mimo-v2.6:9bHugging Face
66
#48
MiniMax M2.5
230B MoE · Q4_K_M · MiniMax · Modified MIT
tok/s
estimated57 M5 Ultra
RAM @64K
estimated146 GB
Min RAM
130 GB
65
#49
Beam 1.0
501.3B MoE · Q4_K_M · Reflection AI · Apache 2.0
tok/s
estimated27 M5 Ultra
RAM @64K
estimated202 GB
Min RAM
199 GB
65
#50
GLM-5.1
754B MoE · Q4_K_M · Zhipu AI · MIT
tok/s
estimated17 M5 Ultra
RAM @64K
estimated396 GB
Min RAM
390 GB
65
#51
Ling-3.0-tiny
7.9B MoE · oQ4e · Ant Group · MIT
tok/s
community174
RAM @64K
estimated7 GB
Min RAM
6 GB
communityprefill 2,012 @4KcommunityTTFT 2.035 s @4K
65
#52
MiniMax-M3
428B MoE · Q4_K_M · MiniMax · MiniMax Community
tok/s
estimated42 M5 Ultra
RAM @64K
estimated170 GB
Min RAM
162 GB
65
#53
Muse Spark Open
816B MoE · oQ3e · Meta · Muse Model License
tok/s
estimated31 M5 Ultra
RAM @64K
—
Min RAM
342 GB
65
#54
Nemotron 3.5 Bolt E4B
4B · Q4_K_M · NVIDIA · OpenMDW
tok/s
estimated161
RAM @64K
estimated8 GB
Min RAM
6 GB
ollama run nemotron-3.5-bolt:e4bHugging Face
65
#55
Qwen 3 30B-A3B
30B MoE · 4-bit · Alibaba · Apache 2.0
tok/s
sourced152
RAM @64K
estimated25 GB
Min RAM
18 GB
ollama run qwen3:30b-a3bHugging Face
65
#56
Qwen 3.5 397B-A17B
397B MoE · 4-bit · Alibaba · Apache 2.0
tok/s
community62 M5 Ultra
RAM @64K
estimated212 GB
Min RAM
210 GB
communityprefill 1,265 @8K
64
#57
DeepSeek R1 8B
8B · Q4_K_M · DeepSeek · MIT
tok/s
estimated91
RAM @64K
estimated16 GB
Min RAM
6 GB
ollama run deepseek-r1:8bHugging Face
64
#58
Gemma 4 31B
31B · 4-bit · Google · Apache 2.0
tok/s
community26
RAM @64K
estimated30 GB
Min RAM
24 GB
communityprefill 710 @4KcommunityTTFT 1.439 s @1K
ollama run gemma4:31bHugging Face
64
#59
Gemma 4 E4B
4B · Q4_K_M · Google · Apache 2.0
tok/s
estimated161
RAM @64K
estimated8 GB
Min RAM
6 GB
ollama run gemma4:e4bHugging Face
64
#60
Laguna S 2.1
118B MoE · Q4_K_M · Poolside · OpenMDW
tok/s
estimated38
RAM @64K
estimated100 GB
Min RAM
96 GB
64
#61
MiniMax-M3.1-Flash
428B MoE · oQ3e · MiniMax · MiniMax Community
tok/s
community35 M5 Ultra
RAM @64K
estimated197 GB
Min RAM
189 GB
communityprefill 1,110 @4KcommunityTTFT 0.969 s @1K
64
#62
Phi-4 Mini
3.8B · Q4_K_M · Microsoft · MIT
tok/s
estimated168
RAM @64K
estimated12 GB
Min RAM
3 GB
ollama run phi4-mini:3.8bHugging Face
64
#63
Step-3.7-Flash
198B MoE · oQ2e · StepFun · Apache 2.0
tok/s
community48
RAM @64K
estimated69 GB
Min RAM
65 GB
communityprefill 594 @4KcommunityTTFT 1.704 s @1K
63
#64
DeepSeek V4 Pro
1.6T MoE · DeepSeek · MIT
tok/s
—
RAM @64K
estimated851 GB
Min RAM
850 GB
63
#65
Hy4 preview
770B MoE · Q4_K_M · Tencent · Apache 2.0
tok/s
estimated14 M5 Ultra
RAM @64K
estimated298 GB
Min RAM
292 GB
63
#66
LFM3-1.7B
1.73B · Q4_K_M · Liquid AI · LFM Open
tok/s
estimated285
RAM @64K
estimated4 GB
Min RAM
3 GB
ollama run lfm3:1.7bHugging Face
63
#67
Mellum2.1-12B-A2.5B-Thinking
12.15B MoE · Q4_K_M · JetBrains · Apache 2.0
tok/s
estimated154
RAM @64K
estimated11 GB
Min RAM
10 GB
63
#68
MiMo-V2.6-Pro
1.02T MoE · Xiaomi · MIT
tok/s
—
RAM @64K
estimated585 GB
Min RAM
581 GB
63
#69
Qwen3-235B-A22B
235B MoE · 4.5 bpw · Alibaba · Apache 2.0
tok/s
estimated26
RAM @64K
estimated140 GB
Min RAM
128 GB
ollama run qwen3:235b-a22bHugging Face
62
#70
Apertus 1.5 8B
8B · Q4_K_M · Swiss AI · Apache 2.0
tok/s
estimated91
RAM @64K
—
Min RAM
5 GB
62
#71
DeepSeek V3.2-Speciale
685B MoE · Q4_K_M · DeepSeek · MIT
tok/s
estimated18 M5 Ultra
RAM @64K
estimated365 GB
Min RAM
360 GB
62
#72
Nemotron 3.5 Super VL 120B-A12B
120B MoE · Q4_0 · NVIDIA · OpenMDW
tok/s
community36
RAM @64K
estimated73 GB
Min RAM
72 GB
communityprefill 885 @4K
ollama run nemotron-3.5-super:120bHugging Face
62
#73
dots3-note preview
280B MoE · 3.5 bpw · Xiaohongshu (RedNote) · Apache 2.0
tok/s
estimated28
RAM @64K
estimated121 GB
Min RAM
120 GB
61
#74
Gemma 4 E2B
2B · Q4_K_M · Google · Apache 2.0
tok/s
estimated261
RAM @64K
estimated3 GB
Min RAM
2 GB
ollama run gemma4:e2bHugging Face
61
#75
Mistral 7B
7B · Q4_K_M · Mistral · Apache 2.0
tok/s
estimated102
RAM @64K
32K max
Min RAM
5 GB
ollama run mistral:7bHugging Face
61
#76
Qwen 3 8B
8B · Q4_K_M · Alibaba · Apache 2.0
tok/s
estimated91
RAM @64K
estimated16 GB
Min RAM
6 GB
ollama run qwen3:8bHugging Face
61
#77
Qwen 3.5 9B
9B · Q4_K_M · Alibaba · Apache 2.0
tok/s
estimated81
RAM @64K
estimated10 GB
Min RAM
7 GB
ollama run qwen3.5:9bHugging Face
60
#78
Granite 4.2 8B
8.79B · Q4_K_M · IBM · Apache 2.0
tok/s
estimated84
RAM @64K
estimated19 GB
Min RAM
8 GB
ollama run granite4.2:8bHugging Face
60
#79
Hunyuan Hy3
295B MoE · Q4_K_M · Tencent · Apache 2.0
tok/s
estimated30 M5 Ultra
RAM @64K
estimated181 GB
Min RAM
160 GB
60
#80
LFM2.5-2.6B
2.6B · Q4_K_M · Liquid AI · LFM Open
tok/s
sourced220
RAM @64K
estimated4 GB
Min RAM
2 GB
sourcedTTFT 0.1 s
60
#81
Ministral 8B
8B · Q4_K_M · Mistral · Apache 2.0
tok/s
estimated91
RAM @64K
estimated12 GB
Min RAM
6 GB
60
#82
Qwen 3.5 122B-A10B
122B MoE · Q4_K_M · Alibaba · Apache 2.0
tok/s
community43
RAM @64K
estimated67 GB
Min RAM
65 GB
ollama run qwen3.5:122b-a10bHugging Face
60
#83
Qwen 3.5 4B
4B · Q4_K_M · Alibaba · Apache 2.0
tok/s
estimated161
RAM @64K
estimated6 GB
Min RAM
3 GB
ollama run qwen3.5:4bHugging Face
59
#84
DeepSeek V3.2
685B MoE · Q4_K_M · DeepSeek · MIT
tok/s
estimated18 M5 Ultra
RAM @64K
estimated365 GB
Min RAM
360 GB
59
#85
Mistral Small 4
119B MoE · Q4_K_M · Mistral · Apache 2.0
tok/s
estimated46
RAM @64K
estimated66 GB
Min RAM
64 GB
59
#86
Qwen 3 4B
4B · Q4_K_M · Alibaba · Apache 2.0
tok/s
estimated161
RAM @64K
estimated13 GB
Min RAM
3 GB
ollama run qwen3:4bHugging Face
59
#87
Ripple-65B Preview
64.8B · Q4_K_M · Ricursive Intelligence · Apache 2.0
tok/s
estimated13
RAM @64K
estimated38 GB
Min RAM
33 GB
ollama run ripple-preview:65bHugging Face
58
#88
Inkling
975B MoE · Thinking Machines · Apache 2.0
tok/s
—
RAM @64K
estimated553 GB
Min RAM
550 GB
58
#89
Kimi K3
2.8T MoE · Moonshot AI · Moonshot Custom
tok/s
—
RAM @64K
estimated1562 GB
Min RAM
1560 GB
58
#90
Kimi K3.1 Pro
2.8T MoE · Moonshot AI · Moonshot Custom
tok/s
—
RAM @64K
estimated1565 GB
Min RAM
1563 GB
58
#91
MiniMax-M2.7
228.7B MoE · IQ3_XXS · MiniMax · MiniMax Non-Commercial
tok/s
estimated45
RAM @64K
estimated98 GB
Min RAM
82 GB
58
#92
Mistral Large 4
1.05T MoE · 3.0 bpw · Mistral · Mistral Ecosystem Open-Weight
tok/s
estimated21 M5 Ultra
RAM @64K
estimated388 GB
Min RAM
383 GB
58
#93
Qwen3.8-2.4T-A95B
2.4T MoE · Alibaba · Qwen Custom
tok/s
—
RAM @64K
estimated1316 GB
Min RAM
1310 GB
58
#94
SmolLM3 3B
3B · Q4_K_M · HuggingFace · Apache 2.0
tok/s
estimated199
RAM @64K
estimated7 GB
Min RAM
2 GB
58
#95
Ternary Bonsai 2 9B
9.41B · PQ2_0 · Prism ML · Apache 2.0
tok/s
sourced77 M5 Pro
RAM @64K
estimated8 GB
Min RAM
5 GB
ollama run bonsai-2:9bHugging Face
57
#96
Llama 3.1 8B
8B · 4-bit · Meta · Meta Custom
tok/s
sourced102
RAM @64K
—
Min RAM
5 GB
ollama run llama3.1:8bHugging Face
57
#97
Step 5 Preview
600B MoE · StepFun · StepFun Model License
tok/s
—
RAM @64K
estimated266 GB
Min RAM
261 GB
56
#98
Gemma 3 4B
4B · Q4_K_M · Google · Gemma
tok/s
estimated161
RAM @64K
—
Min RAM
3 GB
ollama run gemma3:4bHugging Face
56
#99
Granite 4.2 30B
29.28B · Q4_K_M · IBM · Apache 2.0
tok/s
estimated28
RAM @64K
estimated37 GB
Min RAM
20 GB
ollama run granite4.2:30bHugging Face
56
#100
Nemotron 3 Ultra 550B-A55B
550B MoE · Q4_K_M · NVIDIA · OpenMDW
tok/s
estimated12 M5 Ultra
RAM @64K
estimated227 GB
Min RAM
226 GB
55
#101
GLM-4.5-Air
106B MoE · Q4_K_M · Zhipu AI · MIT
tok/s
estimated27
RAM @64K
estimated68 GB
Min RAM
56 GB
55
#102
MiMo-V2-Flash
309B MoE · Q4_K_M · Xiaomi · MIT
tok/s
estimated41 M5 Ultra
RAM @64K
estimated167 GB
Min RAM
165 GB
55
#103
Phi-4 14B
14B · Q4_K_M · Microsoft · MIT
tok/s
estimated55
RAM @64K
estimated23 GB
Min RAM
10 GB
ollama run phi4:14bHugging Face
54
#104
DeepSeek R1
671B MoE · Q4_K_M · DeepSeek · MIT
tok/s
estimated18 M5 Ultra
RAM @64K
estimated355 GB
Min RAM
350 GB
ollama run deepseek-r1:671bHugging Face
54
#105
DeepSeek V3
685B MoE · Q4_K_M · DeepSeek · MIT
tok/s
estimated18 M5 Ultra
RAM @64K
estimated365 GB
Min RAM
360 GB
ollama run deepseek-v3:671bHugging Face
54
#106
Qwen 3 14B
14B · Q4_K_M · Alibaba · Apache 2.0
tok/s
estimated55
RAM @64K
estimated21 GB
Min RAM
10 GB
ollama run qwen3:14bHugging Face
53
#107
Llama 4 Maverick
400B MoE · Q4_K_M · Meta · Meta Custom
tok/s
estimated36 M5 Ultra
RAM @64K
—
Min RAM
210 GB
ollama run llama4:maverickHugging Face
53
#108
Mistral Medium 3.5
128B · Q4_K_M · Mistral · Mistral Modified MIT
tok/s
estimated7
RAM @64K
estimated100 GB
Min RAM
77 GB
ollama run mistral-medium-3.5:128bHugging Face
53
#109
Mistral Small 3.2 24B
24B · Q4_K_M · Mistral · Apache 2.0
tok/s
estimated33
RAM @64K
estimated25 GB
Min RAM
14 GB
ollama run mistral-small3.2:24bHugging Face
52
#110
Devstral Small 24B
24B · Q4_K_M · Mistral · Apache 2.0
tok/s
estimated33
RAM @64K
estimated25 GB
Min RAM
14 GB
52
#111
GLM-4.7
355B MoE · Q4_K_M · Zhipu AI · MIT
tok/s
estimated21 M5 Ultra
RAM @64K
estimated204 GB
Min RAM
180 GB
52
#112
Ministral 3 14B
14B · Q4_K_M · Mistral · Apache 2.0
tok/s
estimated55
RAM @64K
estimated21 GB
Min RAM
10 GB
ollama run ministral-3:14bHugging Face
52
#113
Qwen 2.5 14B
14B · Q4_K_M · Alibaba · Apache 2.0
tok/s
estimated55
RAM @64K
estimated23 GB
Min RAM
10 GB
ollama run qwen2.5:14bHugging Face
51
#114
Gemma 3 12B
12B · Q4_K_M · Google · Gemma
tok/s
estimated64
RAM @64K
—
Min RAM
9 GB
ollama run gemma3:12bHugging Face
51
#115
Mistral Large 3
675B MoE · Q4_K_M · Mistral · Apache 2.0
tok/s
estimated16 M5 Ultra
RAM @64K
estimated360 GB
Min RAM
355 GB
51
#116
Solar Open 2 250B
250B MoE · Upstage · Solar License
tok/s
—
RAM @64K
estimated164 GB
Min RAM
160 GB
51
#117
Step-3.5-Flash
196B MoE · Q4_K_M · StepFun · Apache 2.0
tok/s
estimated29
RAM @64K
estimated124 GB
Min RAM
120 GB
50
#118
DeepSeek R1 32B
32B · Q4_K_M · DeepSeek · MIT
tok/s
estimated25
RAM @64K
estimated39 GB
Min RAM
22 GB
ollama run deepseek-r1:32bHugging Face
50
#119
QwQ 32B
32B · Q4_K_M · Alibaba · Apache 2.0
tok/s
estimated25
RAM @64K
32K max
Min RAM
22 GB
ollama run qwq:32bHugging Face
49
#120
DeepSeek R1 70B
70B · Q4_K_M · DeepSeek · MIT
tok/s
estimated12
RAM @64K
estimated65 GB
Min RAM
44 GB
ollama run deepseek-r1:70bHugging Face
49
#121
Qwen 3 32B
32B · 4-bit · Alibaba · Apache 2.0
tok/s
sourced26
RAM @64K
estimated39 GB
Min RAM
22 GB
ollama run qwen3:32bHugging Face
47
#122
Qwen 3.5 27B
27B · Q4_K_M · Alibaba · Apache 2.0
tok/s
estimated30
RAM @64K
estimated23 GB
Min RAM
18 GB
ollama run qwen3.5:27bHugging Face
46
#123
Apertus 1.5 70B
70B · Q4_K_M · Swiss AI · Apache 2.0
tok/s
estimated12
RAM @64K
—
Min RAM
44 GB
46
#124
Gemma 3 27B
27B · Q4_K_M · Google · Gemma
tok/s
estimated30
RAM @64K
—
Min RAM
18 GB
ollama run gemma3:27bHugging Face
45
#125
Llama 4 Scout
109B MoE · Q4_K_M · Meta · Meta Custom
tok/s
estimated19
RAM @64K
—
Min RAM
58 GB
ollama run llama4:scoutHugging Face
45
#126
Qwen 2.5 72B
72B · Q4_K_M · Alibaba · Apache 2.0
tok/s
estimated12
RAM @64K
estimated67 GB
Min RAM
46 GB
ollama run qwen2.5:72bHugging Face
44
#127
Llama 3.1 405B
405B · Q4_K_M · Meta · Meta Custom
tok/s
estimated4 M5 Ultra
RAM @64K
—
Min RAM
240 GB
ollama run llama3.1:405bHugging Face
42
#128
Hermes 4 70B
70B · Q4_K_M · Nous Research · Meta Custom
tok/s
estimated12
RAM @64K
estimated59 GB
Min RAM
38 GB
41
#129
Llama 3.3 70B
70B · 4-bit · Meta · Meta Custom
tok/s
sourced13
RAM @64K
—
Min RAM
44 GB
sourcedprefill 306 @0.5K
ollama run llama3.3:70bHugging Face
41
#130
Mixtral 8x22B
141B MoE · Q4_K_M · Mistral · Apache 2.0
tok/s
estimated9
RAM @64K
estimated103 GB
Min RAM
88 GB
ollama run mixtral:8x22bHugging Face
Speed is single-stream decode on an M5 Max — or on the Mac Studio M5 Ultra, tagged, for models that need more than 128 GB — at 4-bit unless the row names another quant (shown under each model). Each figure is the lower-median measured plain-decode run on that chip where one exists, otherwise the published estimate; the mark says which, and the score uses exactly the figure shown. Prefill and TTFT appear under a speed only where a run measured them. RAM @64K is the minimum RAM, which already covers a short (about 4K-token) context, plus the 16-bit KV cache the rest of a 65,536-token context adds, computed from each model’s config.json (an 8-bit KV cache halves the added part); “agent-ready” means that total fits the Mac and the model’s own window reaches 64K. Speculative-decoding (MTP) speeds never set a rank. Method & math → · Context & memory → · Why MTP isn’t ranked → 1–130 of 130

Own the Mac already? See which of these 130 models fit your exact chip and unified memory →
Still choosing hardware? Three questions and the Mac Advisor names the config → · Every Mac compared for local AI → · Key numbers →

Questions about the ranking

What is the best local LLM for a Mac right now?

According to the LLMCheck index, as of 18 November 2026, MiMo-V2.6-Flash and Qwen 3.6-35B-A3B tie for #1 at LLMCheck Score 81; MiMo-V2.6-Flash is listed first on the alphabetical tie-break. MiMo-V2.6-Flash runs at 77 tok/s community-measured on an M5 Ultra but needs a Mac Studio with at least 192 GB of unified memory. Qwen 3.6-35B-A3B needs only 24 GB and is the best pick on a Mac with 32 GB or less, at 59 tok/s on a Mac mini M6 32 GB (community-measured on an M6). The full ranking of all 130 models is at llmcheck.net/leaderboard.

What is the best local LLM for each amount of RAM?

According to the LLMCheck index, as of 18 November 2026, the best pick for each RAM tier is the best-ranked model that fits the tier, reaches 15 tok/s on the tier's representative Mac and has a standard Mac runtime (the /best-llm/by-ram/ pick), ranking capability against speed: 8 GB Mac: MiniCPM5-2B (Score 69; ~39 tok/s estimated on a MacBook Air M1 8 GB; needs 3 GB); 16 GB Mac: Gemma 4 12B (Score 66; ~17 tok/s estimated on a Mac mini M6 16 GB; needs 9 GB); 24 GB Mac: Recursive Flint (Score 80; ~71 tok/s estimated on a Mac mini M6 24 GB; needs 20 GB); 32 GB Mac: Qwen 3.6-35B-A3B (Score 81; 59 tok/s on a Mac mini M6 32 GB (community-measured on an M6); needs 24 GB); 64 GB Mac: LFM2.5-24B-A2B (Score 76; 114 tok/s on a Mac mini M5 Pro 64 GB (community-measured on an M5 Pro); needs 17 GB); 128 GB Mac: Qwen 3.6-35B-A3B (Score 81; 130 tok/s on a Mac Studio M5 Max 128 GB (community-measured on an M5 Max); needs 24 GB); 192–256 GB Mac: MiMo-V2.6-Flash (Score 81; 77 tok/s on a Mac Studio M5 Ultra 256 GB (community-measured on an M5 Ultra); needs 170 GB); 512 GB Mac: MiMo-V2.6-Flash (Score 81; 77 tok/s on a Mac Studio M5 Ultra 512 GB (community-measured on an M5 Ultra); needs 170 GB). Each tier, with every Mac in it: llmcheck.net/best-llm/by-ram/.

What is the best dense local LLM?

The strongest dense (non-mixture-of-experts) model in the index is Qwen3.8-27B (capability 46/50, 19 GB minimum, 31 tok/s community-measured on an M5 Max at 4-bit). Dense models read every weight for every token, so they decode slower than mixture-of-experts models of similar size.

What is the fastest local LLM on a Mac?

The fastest measured figure in the index is Maple Preview 20B-A1B at 387 tok/s on an M5 Pro (published, Ternary); its ternary weights decode through compute-bound kernels, so among bandwidth-bound models the fastest measured reference figure is LFM2.5-2.6B at 220 tok/s published on an M5 Max (Q4_K_M).

How is the LLMCheck Score calculated?

Capability (0-50, normalized from published benchmarks) + speed (0-25: the displayed tok/s x 0.25, capped) + RAM accessibility (0-15: 15 at 8 GB or less, down to 3 above 64 GB) + licence openness (0-10: MIT 10 down to non-commercial 2). The speed figure is the lower-median measured plain-decode run on an M5 Max (or an M5 Ultra for models above 128 GB) where one exists, otherwise the published bandwidth estimate. Method: llmcheck.net/methodology#reference-row.

Why did models move in the 19 November 2026 update?

The 19 November 2026 update (data of 18 November) changed no ranking rule. It added a RAM-at-64K figure for each model whose KV-cache layout has been read from its config.json (minimum RAM covers about the first 4K tokens of context; a longer context adds its cache), decode-at-depth and prefill figures where a run measured them, and speculative-decoding (MTP) figures kept beside their paired plain run. None of these enters the Score. As of 18 November 2026, 320 measured rows cover 61 of the 130 models (all listed at llmcheck.net/benchmarks#measured); for 35 of the 130 models the leaderboard figure itself is measured, and the rest show an estimate on their reference chip. The largest gaps between a measurement and the formula's own estimate at the same quant: Nemotron 3.5 Bolt 74B-A9B 64 tok/s community-measured vs 30 estimated on an M5 Max (5-bit); MiMo-V2.6-Flash 77 tok/s community-measured vs 42 estimated on an M5 Ultra (MXFP4); Qwen 3.5 397B-A17B 62 tok/s community-measured vs 36 estimated on an M5 Ultra (4-bit). GLM-4.7-Flash's M5 Max figure (~107 tok/s, estimated) is scaled from its own community-measured M6 run of 46.6 tok/s. Every change is logged at llmcheck.net/changelog.

How much memory does a 64K context need?

According to the LLMCheck index, a model's RAM at 64K is its minimum RAM, which covers about the first 4K tokens, plus the 16-bit KV cache the rest of a 65,536-token context adds, computed from the attention layout in its config.json: an estimate, not a measurement. Qwen 3.6-35B-A3B needs 24 GB at a short context and about 26 GB at 64K (25 GB with an 8-bit KV cache). Sliding-window and linear-attention layers add less. Of the 117 with a recorded layout, 51 are agent-ready at 64K on a 32 GB Mac. Method: llmcheck.net/methodology#context.

Cite this page

LLMCheck. (2026). The local LLM leaderboard. llmcheck.net. https://llmcheck.net/leaderboard (updated ). CC BY 4.0.

How to cite LLMCheck →Data: CC BY 4.0Markdown version