Apple M3
Figures assume the 16 GB unified memory pool shared between CPU and GPU. "What runs on it" is judged at a 8,192-token context. Speeds are estimates, not measurements.
| Model | Sweet-spot quant | Est. speed | Community |
|---|---|---|---|
| DeepSeek-R1-Distill-Llama-8B deepseek-ai | Q8_0 Runs fully on GPU @ 8K ctx | 6–8 tok/s (estimate) | no community data |
| DeepSeek-R1-Distill-Qwen-1.5B deepseek-ai | Q8_0 Runs fully on GPU @ 8K ctx | 28–38 tok/s (estimate) | no community data |
| DeepSeek-R1-Distill-Qwen-14B deepseek-ai | IQ4_XS Runs fully on GPU @ 8K ctx | 6–8 tok/s (estimate) | no community data |
| DeepSeek-R1-Distill-Qwen-7B deepseek-ai | Q8_0 Runs fully on GPU @ 8K ctx | 7–9 tok/s (estimate) | no community data |
| Kimi-VL-A3B-Instruct moonshotai | IQ4_XS Runs fully on GPU @ 8K ctx | no community data | |
| Llama-3.1-8B-Instruct meta-llama | Q8_0 Runs fully on GPU @ 8K ctx | 6–8 tok/s (estimate) | no community data |
| Phi-4-mini-instruct microsoft | Q8_0 Runs fully on GPU @ 8K ctx | 12–16 tok/s (estimate) | no community data |
| Phi-4-reasoning microsoft | IQ4_XS Runs fully on GPU @ 8K ctx | 6–8 tok/s (estimate) | no community data |
| Qwen2.5-7B-Instruct Qwen | Q8_0 Runs fully on GPU @ 8K ctx | 7–9 tok/s (estimate) | no community data |
| Qwen3-14B Qwen | IQ4_XS Runs fully on GPU @ 8K ctx | 6–8 tok/s (estimate) | no community data |
| Qwen3-8B Qwen | Q8_0 Runs fully on GPU @ 8K ctx | 6–8 tok/s (estimate) | no community data |
| Qwen3-Embedding-0.6B Qwen | Q8_0 Runs fully on GPU @ 8K ctx | 38–51 tok/s (estimate) | no community data |
| Qwen3-Embedding-4B Qwen | Q4_K_M Runs fully on GPU @ 8K ctx | 16–22 tok/s (estimate) | no community data |
| Qwen3-Embedding-8B Qwen | Q4_K_M Runs fully on GPU @ 8K ctx | 10–14 tok/s (estimate) | no community data |
| Qwen3-Reranker-0.6B Qwen | BF16 Runs fully on GPU @ 8K ctx | 28–38 tok/s (estimate) | no community data |
| Qwen3-Reranker-4B Qwen | BF16 Runs fully on GPU @ 8K ctx | 6–9 tok/s (estimate) | no community data |
| SmolLM3-3B HuggingFaceTB | BF16 Runs fully on GPU @ 8K ctx | 9–12 tok/s (estimate) | no community data |
| dots.ocr rednote-hilab | BF16 Runs fully on GPU @ 8K ctx | 10–13 tok/s (estimate) | no community data |
| gemma-4-12B-it google | Q4_K_M Runs fully on GPU @ 8K ctx | 8–10 tok/s (estimate) | no community data |
| gemma-4-E2B-it google | Q8_0 Runs fully on GPU @ 8K ctx | 12–16 tok/s (estimate) | no community data |
| gemma-4-E4B-it google | Q8_0 Runs fully on GPU @ 8K ctx | 7–10 tok/s (estimate) | no community data |
| phi-4 microsoft | IQ4_XS Runs fully on GPU @ 8K ctx | 6–8 tok/s (estimate) | no community data |
"Est. speed" is a modelled range labelled estimate (D8) for generation (decode) throughput. "Community" shows the median of approved user-submitted reports on this GPU class only where enough exist — never an estimate. "pp" is measured prompt-processing (ingestion) throughput from approved community reports; rows without a measurement show none.
On this GPU, 12 catalog models run fully on the GPU at an 8,192-token context. The most capable is Kimi-VL-A3B-Instruct at IQ4_XS (needs ~9.7 GiB). Pick a smaller model or a lower quant for more headroom.
Biggest model: Kimi-VL-A3B-Instruct at IQ4_XS
llama-server -m Kimi-VL-A3B-Instruct.i1-IQ4_XS.gguf -c 8192 -ngl 999Derived from the fit engine at an 8,192-token context. See more answer packs.
selected to compare · pick at least 2