AI Index

Open-weight models that run on Apple M4 · 16 GB

  • Unified memory 16 GB
  • Bandwidth 120 GB/s
  • Usable 10.7 GB

Specs: Memory · MacBook Air (13-inch, M4, 2025) - Tech Specs - Apple Support · Bandwidth · MacBook Air (13-inch, M4, 2025) - Tech Specs - Apple Support

FitModelParameters / activePrecisionMemoryEst. speedLongest contextRuns as
Runs wellpplx-embed-v2-context-9b Perplexity8.4BIQ4_XS4.9 / 10.7 GB13–19 tok/s192KUnified memory
Runs wellGemma 4 E4B Google8BQ4_04.8 / 10.7 GB13–19 tok/s125KUnified memory
Runs wellLing-3.0-tiny Ant Group7.9B / 1.3BQ8_0maker’s file8.5 / 10.7 GB37–56 tok/s88KUnified memory
Runs wellQwen2.5-7B-Instruct-1M Alibaba Qwen7.6BQ4_04.9 / 10.7 GB13–19 tok/s113KUnified memory
Runs wellQwen2.5-7B-Instruct Alibaba Qwen7.6BIQ4_XS4.7 / 10.7 GB13–20 tok/s32KUnified memory
Runs wellDeepSeek-R1-Distill-Qwen-7B DeepSeek7.6BQ4_04.9 / 10.7 GB13–19 tok/s32KUnified memory
Runs wellGemma 4 E2B Google5.1BQ6_K4.5 / 10.7 GB14–21 tok/s125KUnified memory
Runs wellWeVisOne-VL-4B Tencent4.8BQ5_K_M4.8 / 10.7 GB13–19 tok/s49KUnified memory
Runs wellQianfan-OCR Baidu4.7BQ5_K_M4.8 / 10.7 GB13–20 tok/s31KUnified memory
Runs wellQwen3.5-4B Alibaba Qwen4.7BQ6_K4.3 / 10.7 GB15–22 tok/s211KUnified memory
Runs wellQwen3-VL-4B-Instruct Alibaba Qwen4.4BQ5_K_M4.6 / 10.7 GB14–21 tok/s51KUnified memory
Runs wellQwen3-4B-Instruct-2507 Alibaba Qwen4BQ6_K4.7 / 10.7 GB13–20 tok/s50KUnified memory
Runs wellQwen3-4B Alibaba Qwen4BQ6_Kmaker’s file4.7 / 10.7 GB13–20 tok/s128KUnified memory
Runs wellDeepSeek-OCR DeepSeek3.3BBF16 (native)maker’s file7.2 / 10.7 GB22–32 tok/s8KUnified memory
Runs wellGranite-4.0-H-Micro IBM3.2BQ8_0maker’s file3.7 / 10.7 GB17–26 tok/s128KUnified memory
Runs wellQwen3-1.7B Alibaba Qwen2BMLX 8bitmaker’s file3.0 / 10.7 GB22–33 tok/s40KUnified memory
Runs wellDeepSeek-R1-Distill-Qwen-1.5B DeepSeek1.8BBF16 (native)maker’s file4.0 / 10.7 GB16–24 tok/s125KUnified memory
Runs wellGLM-OCR Zhipu AI1.3BBF16 (native)maker’s file3.5 / 10.7 GB19–28 tok/s123KUnified memory
Runs wellQwen3-0.6B Alibaba Qwen0.8BBF16 (native)maker’s file2.8 / 10.7 GB25–37 tok/s40KUnified memory
Runs wellEmbeddingGemma 2 Google0.7BBF16 (native)maker’s file2.0 / 10.7 GB38–57 tok/s256KUnified memory
Runs wellQwen3-Embedding-0.6B Alibaba Qwen0.6BBF16 (native)maker’s file2.5 / 10.7 GB28–42 tok/s32KUnified memory
Runs wellbge-reranker-v2-m3 BAAI0.6BFP32 (native)maker’s file3.4 / 10.7 GB20–29 tok/s8KUnified memory
Runs wellbge-m3-unsupervised BAAI0.6BFP32 (native)maker’s file3.4 / 10.7 GB20–29 tok/s8KUnified memory
Runs wellbge-large-en-v1.5 BAAI0.3BFP32 (native)maker’s file1.8 / 10.7 GB43–65 tok/s512Unified memory
Runs wellbge-reranker-base BAAI0.3BFP32 (native)maker’s file1.6 / 10.7 GB53–80 tok/s514Unified memory
Runs wellbge-base-en-v1.5 BAAI0.1BFP32 (native)maker’s file0.9 / 10.7 GB131–197 tok/s512Unified memory
Runs wellbge-small-en-v1.5 BAAI0BFP32 (native)maker’s file0.6 / 10.7 GB420–630 tok/s512Unified memory
Runs, slowlyQwen2.5-14B-Instruct-1M Alibaba Qwen14.8BIQ4_XS9.3 / 10.7 GB6–10 tok/s15KUnified memory
Runs, slowlyQwen2.5-14B-Instruct Alibaba Qwen14.8BIQ4_XS9.3 / 10.7 GB6–10 tok/s15KUnified memory
Runs, slowlyDeepSeek-R1-Distill-Qwen-14B DeepSeek14.8BIQ4_XS9.3 / 10.7 GB6–10 tok/s15KUnified memory
Runs, slowlyQwen3-14B Alibaba Qwen14.8BMLX 4bitmaker’s file9.1 / 10.7 GB7–10 tok/s18KUnified memory
Runs, slowlyGemma 3 12B IT Google12.2BIQ4_XS9.3 / 10.7 GB6–9 tok/s11KUnified memory
Runs, slowlyQwen2.5-Omni-7B Alibaba Qwen10.7BQ6_K9.1 / 10.7 GB6–10 tok/s32KUnified memory
Runs, slowlyGLM-4.6V-Flash Zhipu AI10.3BQ6_K8.7 / 10.7 GB7–10 tok/s59KUnified memory
Runs, slowlyQwen3.5-9B Alibaba Qwen9.7BQ6_K8.1 / 10.7 GB7–11 tok/s89KUnified memory
Runs, slowlyQianfan-VL-8B Baidu8.8BQ6_K8.2 / 10.7 GB7–11 tok/s27KUnified memory
Runs, slowlyGranite Switch 4.1 8B IBM8.8BQ6_Kmaker’s file9.3 / 10.7 GB6–9 tok/s16KUnified memory
Runs, slowlyQwen3-VL-8B-Instruct Alibaba Qwen8.8BQ6_K8.3 / 10.7 GB7–11 tok/s24KUnified memory
Runs, slowlyQwen3-VL-8B-Thinking Alibaba Qwen8.8BQ6_K8.3 / 10.7 GB7–11 tok/s24KUnified memory
Runs, slowlyQwen2.5-VL-7B-Instruct Alibaba Qwen8.3BQ8_09.1 / 10.7 GB6–10 tok/s35KUnified memory
Runs, slowlyQwen3-8B Alibaba Qwen8.2BMLX 8bitmaker’s file9.5 / 10.7 GB6–9 tok/s16KUnified memory
Runs, slowlyDeepSeek-R1-Distill-Llama-8B DeepSeek8BQ8_09.4 / 10.7 GB6–9 tok/s17KUnified memory
Runs, slowlyMistral-7B-Instruct-v0.1 Mistral AI7.2BQ8_08.7 / 10.7 GB7–10 tok/s24KUnified memory

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