AI Index

Open-weight models that run on Apple M5 Max (32-core GPU) · 36 GB

  • Unified memory 36 GB
  • Bandwidth 460 GB/s
  • Usable 24.0 GB

Specs: Memory · MacBook Pro - Tech Specs - Apple · Bandwidth · MacBook Pro - Tech Specs - Apple

FitModelParameters / activePrecisionMemoryEst. speedLongest contextRuns as
Runs wellQwen3.5-35B-A3B Alibaba Qwen36B / 3BQ4_K_M21.5 / 24.0 GB116–173 tok/s135KUnified memory
Runs wellQwen3.6-35B-A3B Alibaba Qwen36B / 3BQ4_K_M21.5 / 24.0 GB116–173 tok/s135KUnified memory
Runs wellQwen3-Omni-30B-A3B-Captioner Alibaba Qwen31.7BQ4_K_M19.6 / 24.0 GB65–97 tok/s55KUnified memory
Runs wellGLM-4.7-Flash Zhipu AI31.2BQ5_021.4 / 24.0 GB57–86 tok/s58KUnified memory
Runs wellQwen3-VL-30B-A3B-Instruct Alibaba Qwen31.1BQ4_K_Mmaker’s file18.9 / 24.0 GB73–110 tok/s62KUnified memory
Runs wellQwen3-VL-30B-A3B-Thinking Alibaba Qwen31.1BQ4_K_Mmaker’s file18.9 / 24.0 GB73–110 tok/s62KUnified memory
Runs wellQwen3-30B-A3B-Instruct-2507 Alibaba Qwen30.5B / 3.3BQ5_021.3 / 24.0 GB75–112 tok/s37KUnified memory
Runs wellQwen3-30B-A3B-Thinking-2507 Alibaba Qwen30.5B / 3.3BQ5_021.3 / 24.0 GB75–112 tok/s37KUnified memory
Runs wellQwen3-30B-A3B Alibaba Qwen30.5B / 3.3BQ5_0maker’s file21.4 / 24.0 GB75–112 tok/s36KUnified memory
Runs wellQwen3-Coder-30B-A3B-Instruct Alibaba Qwen30.5B / 3.3BQ5_021.3 / 24.0 GB75–112 tok/s37KUnified memory
Runs wellHy-Memory-30B-A3B Tencent30.1B / 3BQ5_021.0 / 24.0 GB80–120 tok/s40KUnified memory
Runs wellGranite Switch 4.1 30B IBM30BIQ4_XS17.6 / 24.0 GB13–19 tok/s32KUnified memory
Runs wellQwen3.5-27B Alibaba Qwen27.8BQ4_K_M17.0 / 24.0 GB13–20 tok/s120KUnified memory
Runs wellQwen3.6-27B Alibaba Qwen27.8BQ4_K_M17.0 / 24.0 GB13–20 tok/s120KUnified memory
Runs wellQwen3.8-27B Alibaba Qwen27.8BQ4_K_M17.0 / 24.0 GB13–20 tok/s120KUnified memory
Runs wellGemma 4 26B A4B IT Google25.8B / 3.8BQ6_K21.2 / 24.0 GB63–94 tok/s79KUnified memory
Runs wellMagistral Small 1.2 Mistral AI24BQ5_K_Mmaker’s file17.6 / 24.0 GB13–19 tok/s48KUnified memory
Runs wellDevstral Small 1.0 Mistral AI23.6BQ5_K_M17.6 / 24.0 GB13–19 tok/s48KUnified memory
Runs wellDevstral Small 1.1 Mistral AI23.6BQ5_K_M17.6 / 24.0 GB13–19 tok/s48KUnified memory
Runs wellMagistral Small 1.0 Mistral AI23.6BQ5_K_M17.6 / 24.0 GB13–19 tok/s39KUnified memory
Runs wellMagistral Small 1.1 Mistral AI23.6BQ5_K_Mmaker’s file17.6 / 24.0 GB13–19 tok/s39KUnified memory
Runs wellERNIE-4.5-21B-A3B-Thinking Baidu21.8B / 3BQ6_K18.0 / 24.0 GB78–118 tok/s118KUnified memory
Runs wellGPT OSS 20B OpenAI20.9B / 3.6BMXFP4 (native)maker’s file13.6 / 24.0 GB89–134 tok/s128KUnified memory
Runs wellQwen2.5-14B-Instruct-1M Alibaba Qwen14.8BQ8_016.8 / 24.0 GB13–20 tok/s46KUnified memory
Runs wellQwen2.5-14B-Instruct Alibaba Qwen14.8BQ8_0maker’s file16.9 / 24.0 GB13–20 tok/s32KUnified memory
Runs wellDeepSeek-R1-Distill-Qwen-14B DeepSeek14.8BQ8_016.8 / 24.0 GB13–20 tok/s32KUnified memory
Runs wellQwen3-14B Alibaba Qwen14.8BMLX 8bitmaker’s file16.1 / 24.0 GB14–21 tok/s128KUnified memory
Runs wellGemma 3 12B IT Google12.2BQ8_015.5 / 24.0 GB14–22 tok/s32KUnified memory
Runs wellQwen2.5-Omni-7B Alibaba Qwen10.7BQ8_011.6 / 24.0 GB19–29 tok/s32KUnified memory
Runs wellGLM-4.6V-Flash Zhipu AI10.3BQ8_011.0 / 24.0 GB20–31 tok/s125KUnified memory
Runs wellQwen3.5-9B Alibaba Qwen9.7BQ8_010.3 / 24.0 GB22–33 tok/s256KUnified memory
Runs wellQianfan-VL-8B Baidu8.8BQ8_010.2 / 24.0 GB22–33 tok/s32KUnified memory
Runs wellGranite Switch 4.1 8B IBM8.8BQ8_0maker’s file11.2 / 24.0 GB20–30 tok/s86KUnified memory
Runs wellQwen3-VL-8B-Instruct Alibaba Qwen8.8BQ8_0maker’s file9.7 / 24.0 GB23–35 tok/s109KUnified memory
Runs wellQwen3-VL-8B-Thinking Alibaba Qwen8.8BQ8_0maker’s file9.7 / 24.0 GB23–35 tok/s109KUnified memory
Runs wellpplx-embed-v2-context-9b Perplexity8.4BQ8_09.1 / 24.0 GB25–38 tok/s256KUnified memory
Runs wellQwen2.5-VL-7B-Instruct Alibaba Qwen8.3BBF16 (native)maker’s file16.7 / 24.0 GB13–20 tok/s125KUnified memory
Runs wellQwen3-8B Alibaba Qwen8.2BBF16 (native)maker’s file17.1 / 24.0 GB13–20 tok/s128KUnified memory
Runs wellDeepSeek-R1-Distill-Llama-8B DeepSeek8BBF16 (native)maker’s file16.7 / 24.0 GB13–20 tok/s66KUnified memory
Runs wellGemma 4 E4B Google8BBF16 (native)maker’s file15.8 / 24.0 GB14–21 tok/s125KUnified memory
Runs wellLing-3.0-tiny Ant Group7.9B / 1.3BBF16 (native)maker’s file15.6 / 24.0 GB81–122 tok/s128KUnified memory
Runs wellQwen2.5-7B-Instruct-1M Alibaba Qwen7.6BBF16 (native)maker’s file15.3 / 24.0 GB15–22 tok/s166KUnified memory
Runs wellQwen2.5-7B-Instruct Alibaba Qwen7.6BBF16 (native)maker’s file15.3 / 24.0 GB15–22 tok/s32KUnified memory
Runs wellDeepSeek-R1-Distill-Qwen-7B DeepSeek7.6BBF16 (native)maker’s file15.3 / 24.0 GB15–22 tok/s32KUnified memory
Runs wellMistral-7B-Instruct-v0.1 Mistral AI7.2BBF16 (native)maker’s file15.2 / 24.0 GB15–22 tok/s32KUnified memory
Runs wellGemma 4 E2B Google5.1BBF16 (native)maker’s file10.1 / 24.0 GB22–33 tok/s125KUnified memory
Runs wellWeVisOne-VL-4B Tencent4.8BBF16 (native)maker’s file10.6 / 24.0 GB21–32 tok/s103KUnified memory
Runs wellQianfan-OCR Baidu4.7BBF16 (native)maker’s file10.5 / 24.0 GB22–32 tok/s31KUnified memory
Runs wellQwen3.5-4B Alibaba Qwen4.7BBF16 (native)maker’s file9.4 / 24.0 GB24–36 tok/s256KUnified memory
Runs wellQwen3-VL-4B-Instruct Alibaba Qwen4.4BBF16 (native)maker’s file9.9 / 24.0 GB23–34 tok/s108KUnified memory
Runs wellQwen3-4B-Instruct-2507 Alibaba Qwen4BBF16 (native)maker’s file9.1 / 24.0 GB25–37 tok/s114KUnified memory
Runs wellQwen3-4B Alibaba Qwen4BBF16 (native)maker’s file9.1 / 24.0 GB25–37 tok/s128KUnified memory
Runs wellDeepSeek-OCR DeepSeek3.3BBF16 (native)maker’s file7.2 / 24.0 GB83–124 tok/s8KUnified memory
Runs wellGranite-4.0-H-Micro IBM3.2BBF16 (native)maker’s file6.5 / 24.0 GB36–53 tok/s128KUnified memory
Runs wellQwen3-1.7B Alibaba Qwen2BBF16 (native)maker’s file5.2 / 24.0 GB46–69 tok/s40KUnified memory
Runs wellDeepSeek-R1-Distill-Qwen-1.5B DeepSeek1.8BBF16 (native)maker’s file4.0 / 24.0 GB61–91 tok/s125KUnified memory
Runs wellGLM-OCR Zhipu AI1.3BBF16 (native)maker’s file3.5 / 24.0 GB72–108 tok/s128KUnified memory
Runs wellQwen3-0.6B Alibaba Qwen0.8BBF16 (native)maker’s file2.8 / 24.0 GB94–141 tok/s40KUnified memory
Runs wellEmbeddingGemma 2 Google0.7BBF16 (native)maker’s file2.0 / 24.0 GB146–219 tok/s256KUnified memory
Runs wellQwen3-Embedding-0.6B Alibaba Qwen0.6BBF16 (native)maker’s file2.5 / 24.0 GB108–162 tok/s32KUnified memory
Runs wellbge-reranker-v2-m3 BAAI0.6BFP32 (native)maker’s file3.4 / 24.0 GB75–112 tok/s8KUnified memory
Runs wellbge-m3-unsupervised BAAI0.6BFP32 (native)maker’s file3.4 / 24.0 GB75–112 tok/s8KUnified memory
Runs wellbge-large-en-v1.5 BAAI0.3BFP32 (native)maker’s file1.8 / 24.0 GB165–248 tok/s512Unified memory
Runs wellbge-reranker-base BAAI0.3BFP32 (native)maker’s file1.6 / 24.0 GB203–305 tok/s514Unified memory
Runs wellbge-base-en-v1.5 BAAI0.1BFP32 (native)maker’s file0.9 / 24.0 GB503–755 tok/s512Unified memory
Runs wellbge-small-en-v1.5 BAAI0BFP32 (native)maker’s file0.6 / 24.0 GB1609–2414 tok/s512Unified memory
Runs, slowlyQwen2.5-VL-32B-Instruct Alibaba Qwen33.5BQ4_020.4 / 24.0 GB11–16 tok/s22KUnified memory
Runs, slowlyQwen3-VL-32B-Instruct Alibaba Qwen33.4BQ4_K_Mmaker’s file21.3 / 24.0 GB10–16 tok/s18KUnified memory
Runs, slowlyQwen3-VL-32B-Thinking Alibaba Qwen33.4BQ4_K_Mmaker’s file21.3 / 24.0 GB10–16 tok/s18KUnified memory
Runs, slowlyQwen2.5-32B-Instruct Alibaba Qwen32.8BQ4_K_Mmaker’s file21.4 / 24.0 GB10–16 tok/s18KUnified memory
Runs, slowlyQwen2.5-Coder-32B-Instruct Alibaba Qwen32.8BINT4 (AWQ / GPTQ)maker’s file20.9 / 24.0 GB11–16 tok/s20KUnified memory
Runs, slowlyQwQ-32B Alibaba Qwen32.8BQ4_K_Mmaker’s file21.4 / 24.0 GB10–16 tok/s18KUnified memory
Runs, slowlyDeepSeek-R1-Distill-Qwen-32B DeepSeek32.8BQ4_K_M21.5 / 24.0 GB10–16 tok/s18KUnified memory
Runs, slowlyQwen3-32B Alibaba Qwen32.8BQ4_K_Mmaker’s file21.3 / 24.0 GB10–16 tok/s18KUnified memory
Runs, slowlyGemma 4 31B IT Google31.3BQ4_K_M20.6 / 24.0 GB11–16 tok/s30KUnified memory

Estimates from the makers’ Hugging Face repositories (as of 2026-11-06); filtered views aren’t indexed. How it’s estimated