MacBook Air M3 8GB

macOS Mid tier · Apple M3 · 8 GB RAM · 100 GB/s bandwidth
View device page →
BENCHMARK RESULTS (67)
Model Quant Measured Estimated RAM used App Source Date
Whisper Tiny Whisper Tiny FP16 221 tok/s 200 tok/s +11% 0.1 GB Xybrid CLI community 2026-01-03
SmolLM2 135M SmolLM2 135M FP16 212 tok/s 200 tok/s +6% 0.3 GB Xybrid CLI community 2026-01-03
Wav2Vec2 Base Wav2Vec2 Base FP16 211 tok/s 200 tok/s +6% 0.3 GB MacWhisper editorial 2026-02-27
🔢 all-MiniLM-L6-v2 FP16 207 tok/s 200 tok/s +4% 0.1 GB Jan community 2026-03-12
🔢 BGE Small FP16 203 tok/s 200 tok/s +2% 0.1 GB Ollama community 2026-02-10
🗣️ KittenTTS Nano FP16 197 tok/s 200 tok/s -1% 0.1 GB Piper community 2026-01-28
🗣️ Kokoro 82M FP16 191 tok/s 200 tok/s -4% 0.4 GB Xybrid CLI community 2026-04-02
Whisper Small Whisper Small FP16 186 tok/s 200 tok/s -7% 0.5 GB MacWhisper community 2026-05-28
🔢 Nomic Embed Text FP16 185 tok/s 200 tok/s -7% 0.3 GB Jan community 2026-04-10
Ternary Bonsai 1.7B Ternary Bonsai 1.7B Q2 179 tok/s 200 tok/s -10% 0.5 GB Ollama community 2026-02-23
🗣️ KittenTTS Mini FP16 178 tok/s 200 tok/s -11% 0.2 GB Piper editorial 2026-04-20
🔢 GTE Large FP16 143 tok/s 149 tok/s -4% 0.9 GB LM Studio editorial 2026-01-03
Ternary Bonsai 4B Ternary Bonsai 4B Q2 125 tok/s 116 tok/s +8% 0.9 GB Xybrid CLI community 2026-05-27
SmolLM2 360M SmolLM2 360M FP16 101 tok/s 109 tok/s -7% 1.1 GB LM Studio community 2026-02-10
🗣️ OuteTTS 0.3 500M FP16 98 tok/s 100 tok/s -2% 1.2 GB Xybrid CLI community 2026-04-19
👁️ SmolVLM 500M FP16 91 tok/s 91 tok/s +0% 1.4 GB Jan community 2026-05-27
Qwen 2.5 Coder 0.5B Qwen 2.5 Coder 0.5B FP16 88 tok/s 83 tok/s +6% 1.4 GB Xybrid CLI community 2026-03-12
Qwen 2.5 0.5B Qwen 2.5 0.5B FP16 85 tok/s 83 tok/s +2% 1.5 GB LM Studio editorial 2026-02-24
Bonsai 8B (1-bit) Bonsai 8B (1-bit) Q1 79 tok/s 86 tok/s -8% 1.2 GB Jan community 2026-03-01
🗣️ NeuTTS Air FP16 74 tok/s 67 tok/s +10% 1.7 GB Piper community 2026-04-10
🎙️ Distil-Whisper Large V3 FP16 71 tok/s 67 tok/s +6% 1.6 GB MacWhisper community 2026-02-25
Whisper Medium Whisper Medium FP16 69 tok/s 67 tok/s +3% 1.6 GB MacWhisper community 2026-03-28
Ternary Bonsai 8B Ternary Bonsai 8B Q2 64 tok/s 63 tok/s +2% 2.0 GB Jan community 2026-05-27
Qwen 3.5 0.8B Qwen 3.5 0.8B FP16 60 tok/s 56 tok/s +7% 1.9 GB Ollama community 2026-04-08
Bonsai Image 4B Bonsai Image 4B Q2 50 tok/s 56 tok/s -11% 2.3 GB Jan community 2026-02-20
Gemma 3 1B Gemma 3 1B FP16 45 tok/s 45 tok/s +0% 2.4 GB Jan community 2026-03-12
LFM2.5 1.2B LFM2.5 1.2B FP16 45 tok/s 43 tok/s +5% 3.0 GB Ollama editorial 2026-03-28
🎨 Stable Diffusion Turbo FP16 43 tok/s 48 tok/s -10% 2.4 GB Jan community 2026-05-12
TinyLlama 1.1B TinyLlama 1.1B FP16 40 tok/s 45 tok/s -11% 2.5 GB Jan community 2026-05-12
Llama 3.2 1B Llama 3.2 1B FP16 40 tok/s 37 tok/s +8% 3.3 GB Xybrid CLI community 2026-03-16
Qwen 2.5 Coder 1.5B Qwen 2.5 Coder 1.5B FP16 34 tok/s 31 tok/s +10% 3.9 GB Jan community 2026-01-14
Whisper Large V3 Whisper Large V3 FP16 34 tok/s 32 tok/s +6% 3.4 GB MacWhisper community 2026-04-07
🗣️ Dia 1.6B FP16 33 tok/s 30 tok/s +10% 3.9 GB Xybrid CLI community 2026-03-12
👨‍💻 StarCoder2 3B Q8 30 tok/s 30 tok/s +0% 4.2 GB Jan community 2026-05-25
StableLM 2 1.6B StableLM 2 1.6B FP16 28 tok/s 30 tok/s -7% 3.7 GB Xybrid CLI community 2026-05-12
Qwen 2.5 3B Qwen 2.5 3B Q8 28 tok/s 29 tok/s -3% 3.8 GB Ollama community 2026-01-05
🧠 DeepSeek R1 Distill 1.5B FP16 28 tok/s 30 tok/s -7% 3.9 GB Ollama community 2026-03-18
SmolLM2 1.7B SmolLM2 1.7B FP16 26 tok/s 29 tok/s -10% 4.3 GB Ollama community 2026-02-23
Llama 3.2 3B Llama 3.2 3B Q8 26 tok/s 28 tok/s -7% 4.0 GB LM Studio community 2026-05-23
Qwen 2.5 Coder 3B Qwen 2.5 Coder 3B Q8 26 tok/s 29 tok/s -10% 3.9 GB Jan community 2026-05-11
Qwen 3.5 2B Qwen 3.5 2B FP16 25 tok/s 24 tok/s +4% 4.7 GB Jan community 2026-02-09
👁️ Moondream 2B FP16 25 tok/s 27 tok/s -7% 3.9 GB Jan community 2026-02-23
🎨 SDXL Turbo Q8 25 tok/s 26 tok/s -4% 4.7 GB Xybrid CLI community 2026-05-14
Qwen 3.5 4B Qwen 3.5 4B Q8 24 tok/s 23 tok/s +4% 5.2 GB Jan community 2026-03-14
🎨 Stable Diffusion 3.5 Medium FP16 24 tok/s 22 tok/s +9% 5.1 GB Xybrid CLI community 2026-01-05
Phi-4 Mini Phi-4 Mini Q8 23 tok/s 23 tok/s +0% 5.6 GB Jan community 2026-04-21
Gemma 4 E2B Gemma 4 E2B FP16 22 tok/s 22 tok/s +0% 5.0 GB Xybrid CLI community 2026-05-09
Gemma 4 E4B Gemma 4 E4B Q8 21 tok/s 20 tok/s +5% 5.9 GB LM Studio community 2026-05-08
Mistral 7B Mistral 7B Q5 21 tok/s 20 tok/s +5% 6.3 GB LM Studio editorial 2026-03-02
Phi-4 Medium Phi-4 Medium Q2 21 tok/s 21 tok/s +0% 5.3 GB LM Studio editorial 2026-05-12
👨‍💻 DeepSeek Coder 6.7B Q6 21 tok/s 19 tok/s +11% 5.5 GB Xybrid CLI community 2026-03-18
Gemma 3n E4B Gemma 3n E4B Q5 20 tok/s 20 tok/s +0% 6.2 GB Xybrid CLI community 2026-01-18
🧠 Qwen3 8B Q5 20 tok/s 18 tok/s +11% 6.1 GB Xybrid CLI editorial 2026-01-03
Qwen 2.5 Coder 7B Qwen 2.5 Coder 7B Q5 20 tok/s 20 tok/s +0% 5.4 GB Xybrid CLI community 2026-01-19
👁️ LLaVA 1.6 7B Q6 20 tok/s 18 tok/s +11% 6.2 GB Jan community 2026-05-12
Qwen 2.5 VL 7B Qwen 2.5 VL 7B Q5 20 tok/s 19 tok/s +5% 6.3 GB LM Studio community 2026-05-10
🎨 FLUX.1 Schnell Q3 20 tok/s 20 tok/s +0% 5.4 GB Ollama community 2026-02-06
Gemma 3n E2B Gemma 3n E2B Q8 19 tok/s 21 tok/s -10% 5.5 GB Jan community 2026-05-14
Gemma 3 4B Gemma 3 4B Q8 19 tok/s 20 tok/s -5% 5.8 GB Ollama community 2026-04-23
LFM2.5 8B A1B LFM2.5 8B A1B Q5 19 tok/s 18 tok/s +6% 7.1 GB LM Studio editorial 2026-03-03
🧠 DeepSeek R1 Distill 7B Q5 19 tok/s 20 tok/s -5% 6.0 GB Xybrid CLI community 2026-01-19
Qwen 2.5 7B Qwen 2.5 7B Q5 18 tok/s 20 tok/s -10% 6.5 GB Xybrid CLI editorial 2026-04-18
Llama 3.1 8B Llama 3.1 8B Q5 18 tok/s 19 tok/s -5% 6.8 GB LM Studio editorial 2026-05-25
Mistral Nemo 12B Mistral Nemo 12B Q3 18 tok/s 19 tok/s -5% 6.4 GB Xybrid CLI community 2026-01-17
Gemma 3 12B Gemma 3 12B Q3 17 tok/s 19 tok/s -11% 6.2 GB Jan community 2026-05-16
🧠 DeepSeek R1 Distill 8B Q5 17 tok/s 19 tok/s -11% 6.0 GB LM Studio editorial 2026-03-16
Qwen 3.5 9B Qwen 3.5 9B Q4 16 tok/s 18 tok/s -11% 6.5 GB LM Studio community 2026-03-14
← All benchmarks