AI Index

Open-weight models that run on Apple M5 Pro (16-core GPU) · 24 GB

  • Unified memory 24 GB
  • Bandwidth 307 GB/s
  • Usable 16.0 GB

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

FitModelParameters / activePrecisionMemoryEst. speedLongest contextRuns as
Runs wellGemma 4 26B A4B IT Google25.8B / 3.8BIQ4_XS13.9 / 16.0 GB60–90 tok/s61KUnified memory
Runs wellERNIE-4.5-21B-A3B-Thinking Baidu21.8B / 3BQ4_K_M13.4 / 16.0 GB67–100 tok/s55KUnified memory
Runs wellGPT OSS 20B OpenAI20.9B / 3.6BMXFP4 (native)maker’s file13.6 / 16.0 GB60–89 tok/s108KUnified memory
Runs wellQwen2.5-14B-Instruct-1M Alibaba Qwen14.8BQ5_K_M11.8 / 16.0 GB13–19 tok/s30KUnified memory
Runs wellQwen2.5-14B-Instruct Alibaba Qwen14.8BQ5_K_Mmaker’s file11.8 / 16.0 GB13–19 tok/s30KUnified memory
Runs wellDeepSeek-R1-Distill-Qwen-14B DeepSeek14.8BQ5_K_M11.8 / 16.0 GB13–19 tok/s30KUnified memory
Runs wellQwen3-14B Alibaba Qwen14.8BQ5_K_Mmaker’s file11.5 / 16.0 GB13–19 tok/s36KUnified memory
Runs wellGemma 3 12B IT Google12.2BQ5_K_M11.4 / 16.0 GB13–20 tok/s21KUnified memory
Runs wellQwen2.5-Omni-7B Alibaba Qwen10.7BQ8_011.6 / 16.0 GB13–19 tok/s32KUnified memory
Runs wellGLM-4.6V-Flash Zhipu AI10.3BQ8_011.0 / 16.0 GB14–20 tok/s125KUnified memory
Runs wellQwen3.5-9B Alibaba Qwen9.7BQ8_010.3 / 16.0 GB15–22 tok/s190KUnified memory
Runs wellQianfan-VL-8B Baidu8.8BQ8_010.2 / 16.0 GB15–22 tok/s32KUnified memory
Runs wellGranite Switch 4.1 8B IBM8.8BQ8_0maker’s file11.2 / 16.0 GB13–20 tok/s37KUnified memory
Runs wellQwen3-VL-8B-Instruct Alibaba Qwen8.8BQ8_0maker’s file9.7 / 16.0 GB15–23 tok/s52KUnified memory
Runs wellQwen3-VL-8B-Thinking Alibaba Qwen8.8BQ8_0maker’s file9.7 / 16.0 GB15–23 tok/s52KUnified memory
Runs wellpplx-embed-v2-context-9b Perplexity8.4BQ8_09.1 / 16.0 GB17–25 tok/s230KUnified memory
Runs wellQwen2.5-VL-7B-Instruct Alibaba Qwen8.3BQ8_09.1 / 16.0 GB17–25 tok/s125KUnified memory
Runs wellQwen3-8B Alibaba Qwen8.2BMLX 8bitmaker’s file9.5 / 16.0 GB16–24 tok/s128KUnified memory
Runs wellDeepSeek-R1-Distill-Llama-8B DeepSeek8BQ8_09.4 / 16.0 GB16–24 tok/s60KUnified memory
Runs wellGemma 4 E4B Google8BQ8_08.6 / 16.0 GB18–27 tok/s125KUnified memory
Runs wellLing-3.0-tiny Ant Group7.9B / 1.3BQ8_0maker’s file8.5 / 16.0 GB95–143 tok/s128KUnified memory
Runs wellQwen2.5-7B-Instruct-1M Alibaba Qwen7.6BQ8_08.5 / 16.0 GB18–27 tok/s145KUnified memory
Runs wellQwen2.5-7B-Instruct Alibaba Qwen7.6BQ8_0maker’s file8.5 / 16.0 GB18–27 tok/s32KUnified memory
Runs wellDeepSeek-R1-Distill-Qwen-7B DeepSeek7.6BQ8_08.5 / 16.0 GB18–27 tok/s32KUnified memory
Runs wellMistral-7B-Instruct-v0.1 Mistral AI7.2BQ8_08.7 / 16.0 GB18–26 tok/s32KUnified memory
Runs wellGemma 4 E2B Google5.1BBF16 (native)maker’s file10.1 / 16.0 GB15–22 tok/s125KUnified memory
Runs wellWeVisOne-VL-4B Tencent4.8BBF16 (native)maker’s file10.6 / 16.0 GB14–21 tok/s46KUnified memory
Runs wellQianfan-OCR Baidu4.7BBF16 (native)maker’s file10.5 / 16.0 GB14–22 tok/s31KUnified memory
Runs wellQwen3.5-4B Alibaba Qwen4.7BBF16 (native)maker’s file9.4 / 16.0 GB16–24 tok/s218KUnified memory
Runs wellQwen3-VL-4B-Instruct Alibaba Qwen4.4BBF16 (native)maker’s file9.9 / 16.0 GB15–23 tok/s51KUnified memory
Runs wellQwen3-4B-Instruct-2507 Alibaba Qwen4BBF16 (native)maker’s file9.1 / 16.0 GB17–25 tok/s57KUnified memory
Runs wellQwen3-4B Alibaba Qwen4BBF16 (native)maker’s file9.1 / 16.0 GB17–25 tok/s128KUnified memory
Runs wellDeepSeek-OCR DeepSeek3.3BBF16 (native)maker’s file7.2 / 16.0 GB55–83 tok/s8KUnified memory
Runs wellGranite-4.0-H-Micro IBM3.2BBF16 (native)maker’s file6.5 / 16.0 GB24–36 tok/s128KUnified memory
Runs wellQwen3-1.7B Alibaba Qwen2BBF16 (native)maker’s file5.2 / 16.0 GB31–46 tok/s40KUnified memory
Runs wellDeepSeek-R1-Distill-Qwen-1.5B DeepSeek1.8BBF16 (native)maker’s file4.0 / 16.0 GB41–61 tok/s125KUnified memory
Runs wellGLM-OCR Zhipu AI1.3BBF16 (native)maker’s file3.5 / 16.0 GB48–72 tok/s128KUnified memory
Runs wellQwen3-0.6B Alibaba Qwen0.8BBF16 (native)maker’s file2.8 / 16.0 GB63–94 tok/s40KUnified memory
Runs wellEmbeddingGemma 2 Google0.7BBF16 (native)maker’s file2.0 / 16.0 GB97–146 tok/s256KUnified memory
Runs wellQwen3-Embedding-0.6B Alibaba Qwen0.6BBF16 (native)maker’s file2.5 / 16.0 GB72–108 tok/s32KUnified memory
Runs wellbge-reranker-v2-m3 BAAI0.6BFP32 (native)maker’s file3.4 / 16.0 GB50–75 tok/s8KUnified memory
Runs wellbge-m3-unsupervised BAAI0.6BFP32 (native)maker’s file3.4 / 16.0 GB50–75 tok/s8KUnified memory
Runs wellbge-large-en-v1.5 BAAI0.3BFP32 (native)maker’s file1.8 / 16.0 GB110–166 tok/s512Unified memory
Runs wellbge-reranker-base BAAI0.3BFP32 (native)maker’s file1.6 / 16.0 GB136–204 tok/s514Unified memory
Runs wellbge-base-en-v1.5 BAAI0.1BFP32 (native)maker’s file0.9 / 16.0 GB336–504 tok/s512Unified memory
Runs wellbge-small-en-v1.5 BAAI0BFP32 (native)maker’s file0.6 / 16.0 GB1074–1611 tok/s512Unified memory
Runs, slowlyQwen3.5-27B Alibaba Qwen27.8BQ4_015.8 / 16.0 GB9–14 tok/s11KUnified memory
Runs, slowlyQwen3.6-27B Alibaba Qwen27.8BQ4_015.8 / 16.0 GB9–14 tok/s11KUnified memory
Runs, slowlyQwen3.8-27B Alibaba Qwen27.8BQ4_015.8 / 16.0 GB9–14 tok/s11KUnified memory
Runs, slowlyMagistral Small 1.2 Mistral AI24BIQ4_XS13.7 / 16.0 GB11–16 tok/s22KUnified memory
Runs, slowlyDevstral Small 1.0 Mistral AI23.6BQ4_014.2 / 16.0 GB11–16 tok/s19KUnified memory
Runs, slowlyDevstral Small 1.1 Mistral AI23.6BQ4_014.2 / 16.0 GB11–16 tok/s19KUnified memory
Runs, slowlyMagistral Small 1.0 Mistral AI23.6BQ4_014.2 / 16.0 GB11–16 tok/s19KUnified memory
Runs, slowlyMagistral Small 1.1 Mistral AI23.6BQ4_014.2 / 16.0 GB11–16 tok/s19KUnified memory

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