Best models for this Mac

29 modelsCatalog current through February 27, 2026

Mac Studio M3 Ultra 96GB ranked for coding with a most capable bias, using the best available runtime evidence.

Use the strongest current runtime evidence for each row.M5 Max watch: 8B 61.6 tok/s · 14B 34.3 tok/s on the current 36GB public anchor.
RankModelScoreQuantTok/sRuntimeEvidenceHeadroomWhy it ranks here
1Qwen 3 32B2748bit22.0 tok/sOllamaEstimated63.0 GB8bit is the highest practical quality here. 22.0 tok/s estimated from nearby benchmark coverage, with Ollama wrapper on llama.cpp as the best runtime hint. 63.0 GB headroom leaves workable context margin.
2Qwen3.5-27B2648bit20.6 tok/sMLXEstimated68.4 GB8bit is the highest practical quality here. 20.6 tok/s estimated from nearby benchmark coverage, with MLX backend as the best runtime hint. 68.4 GB headroom leaves workable context margin.
3Devstral Small 1.12628bit43.0 tok/sLM StudioEstimated71.9 GB8bit is the highest practical quality here. 43.0 tok/s estimated from nearby benchmark coverage, with LM Studio wrapper on mixed as the best runtime hint. 71.9 GB headroom leaves workable context margin.
4Devstral Small 2 24B2628bit25.3 tok/sMLXEstimated71.9 GB8bit is the highest practical quality here. 25.3 tok/s estimated from nearby benchmark coverage, with MLX backend as the best runtime hint. 71.9 GB headroom leaves workable context margin.
5Llama 3.3 70B2618bit11.8 tok/sLM StudioEstimated27.2 GB8bit is the highest practical quality here. 11.8 tok/s estimated from nearby benchmark coverage, with LM Studio wrapper on mixed as the best runtime hint. 27.2 GB headroom leaves workable context margin.
6Gemma 3 27B2598bit14.5 tok/sLM StudioEstimated66.1 GB8bit is the highest practical quality here. 14.5 tok/s estimated from nearby benchmark coverage, with LM Studio wrapper on mixed as the best runtime hint. 66.1 GB headroom leaves workable context margin.
7DeepSeek R1 Distill Qwen 32B2528bitMeasure itBest availableFit-first63.0 GB8bit is the highest practical quality here. Speed still needs direct measurement. 63.0 GB headroom leaves workable context margin.
8Mistral Small 3.1 24B2408bitMeasure itBest availableFit-first71.9 GB8bit is the highest practical quality here. Speed still needs direct measurement. 71.9 GB headroom leaves workable context margin.
9Magistral Small2408bitMeasure itBest availableFit-first71.9 GB8bit is the highest practical quality here. Speed still needs direct measurement. 71.9 GB headroom leaves workable context margin.
10Nemotron-3-Nano-30B-A3B2338bit43.7 tok/sllama.cppEstimated67.2 GB8bit is the highest practical quality here. 43.7 tok/s estimated from nearby benchmark coverage, with llama.cpp backend as the best runtime hint. 67.2 GB headroom leaves workable context margin.
11Qwen 3 30B-A3B2338bit76.7 tok/sMLXEstimated66.3 GB8bit is the highest practical quality here. 76.7 tok/s estimated from nearby benchmark coverage, with MLX backend as the best runtime hint. 66.3 GB headroom leaves workable context margin.
12Qwen3-Coder-30B-A3B2338bit58.5 tok/sllama.cppEstimated66.7 GB8bit is the highest practical quality here. 58.5 tok/s estimated from nearby benchmark coverage, with llama.cpp backend as the best runtime hint. 66.7 GB headroom leaves workable context margin.
13Qwen3.5-35B-A3B2328bit80.0 tok/sMLXEstimated62.3 GB8bit is the highest practical quality here. 80.0 tok/s estimated from nearby benchmark coverage, with MLX backend as the best runtime hint. 62.3 GB headroom leaves workable context margin.
14GLM-4.7-Flash2328bit36.8 tok/sllama.cppEstimated60.2 GB8bit is the highest practical quality here. 36.8 tok/s estimated from nearby benchmark coverage, with llama.cpp backend as the best runtime hint. 60.2 GB headroom leaves workable context margin.
15Qwen 3 8B2118bit63.1 tok/sLM StudioEstimated86.7 GB8bit is the highest practical quality here. 63.1 tok/s estimated from nearby benchmark coverage, with LM Studio wrapper on mixed as the best runtime hint. 86.7 GB headroom leaves workable context margin.
16Llama 2 7B2098bit36.4 tok/sllama.cppEstimated85.2 GB8bit is the highest practical quality here. 36.4 tok/s estimated from nearby benchmark coverage, with llama.cpp backend as the best runtime hint. 85.2 GB headroom leaves workable context margin.
17Qwen3.5-9B2068bit15.0 tok/sllama.cppEstimated86.1 GB8bit is the highest practical quality here. 15.0 tok/s estimated from nearby benchmark coverage, with llama.cpp backend as the best runtime hint. 86.1 GB headroom leaves workable context margin.
18Qwen 2.5 14B1998bitMeasure itBest availableFit-first81.0 GB8bit is the highest practical quality here. Speed still needs direct measurement. 81.0 GB headroom leaves workable context margin.
19Phi-4 14B1988bitMeasure itBest availableFit-first81.5 GB8bit is the highest practical quality here. Speed still needs direct measurement. 81.5 GB headroom leaves workable context margin.
20Llama 3.1 8B1898bitMeasure itBest availableFit-first87.0 GB8bit is the highest practical quality here. Speed still needs direct measurement. 87.0 GB headroom leaves workable context margin.
21Qwen 2.5 7B1898bitMeasure itBest availableFit-first88.0 GB8bit is the highest practical quality here. Speed still needs direct measurement. 88.0 GB headroom leaves workable context margin.
22Mistral 7B v0.31888bitMeasure itBest availableFit-first87.7 GB8bit is the highest practical quality here. Speed still needs direct measurement. 87.7 GB headroom leaves workable context margin.
23GLM-4.5-Air1796bit54.0 tok/sMLXEstimated20.0 GB6bit is the highest practical quality here. 54.0 tok/s estimated from nearby benchmark coverage, with MLX backend as the best runtime hint. 20.0 GB headroom leaves workable context margin.
24Qwen3.5-122B-A10B174Q557.0 tok/sMLXEstimated23.7 GBQ5 is the highest practical quality here. 57.0 tok/s estimated from nearby benchmark coverage, with MLX backend as the best runtime hint. 23.7 GB headroom leaves workable context margin.
25Qwen3-Coder-Next1728bit74.0 tok/sMLXEstimated20.2 GB8bit is the highest practical quality here. 74.0 tok/s estimated from nearby benchmark coverage, with MLX backend as the best runtime hint. 20.2 GB headroom leaves workable context margin.
26Gemma 3 4B1508bit100.5 tok/sLM StudioEstimated90.4 GB8bit is the highest practical quality here. 100.5 tok/s estimated from nearby benchmark coverage, with LM Studio wrapper on mixed as the best runtime hint. 90.4 GB headroom leaves workable context margin.
27Qwen 3 4B1508bit143.2 tok/sMLXEstimated90.6 GB8bit is the highest practical quality here. 143.2 tok/s estimated from nearby benchmark coverage, with MLX backend as the best runtime hint. 90.6 GB headroom leaves workable context margin.
28Qwen 3 0.6B1458bit184.4 tok/sLM StudioEstimated94.8 GB8bit is the highest practical quality here. 184.4 tok/s estimated from nearby benchmark coverage, with LM Studio wrapper on mixed as the best runtime hint. 94.8 GB headroom leaves workable context margin.
29Llama 3.2 1B1248bitMeasure itBest availableFit-first94.1 GB8bit is the highest practical quality here. Speed still needs direct measurement. 94.1 GB headroom leaves workable context margin.
96GBUnified memory
$3,999MSRP
mac_studioForm factor
m3-ultraChip