What can your GPU run?
Pick your card or Mac. Each page lists every model it can run, how fast, and the command to start. Choosing between two? See the head-to-head comparisons.
NVIDIA
- GTX 9704 GB · best: Qwen3 4B
- GTX 9804 GB · best: Qwen3 4B
- GTX 980 Ti6 GB · best: Qwen3 4B
- GTX 10606 / 3 GB · best: Qwen3 4B
- GTX 10708 GB · best: Qwen3 8B
- GTX 1070 Ti8 GB · best: Qwen3 8B
- GTX 10808 GB · best: Qwen3 8B
- GTX 1080 Ti11 GB · best: Qwen3 14B
- GTX 16504 GB · best: Qwen3 4B
- GTX 16606 GB · best: Qwen3 4B
- GTX 1660 Super6 GB · best: Qwen3 4B
- GTX 1660 Ti6 GB · best: Qwen3 4B
- RTX 20606 / 12 GB · best: Qwen3 4B
- RTX 2060 Super8 GB · best: Qwen3 8B
- RTX 20708 GB · best: Qwen3 8B
- RTX 2070 Super8 GB · best: Qwen3 8B
- RTX 20808 GB · best: Qwen3 8B
- RTX 2080 Super8 GB · best: Qwen3 8B
- RTX 2080 Ti11 GB · best: Qwen3 14B
- RTX 30508 / 6 GB · best: Qwen3 8B
- RTX 306012 / 8 GB · best: Qwen3 14B
- RTX 3060 Ti8 GB · best: Qwen3 8B
- RTX 30708 GB · best: Qwen3 8B
- RTX 3070 Ti8 GB · best: Qwen3 8B
- RTX 308010 / 12 GB · best: Gemma 3 12B
- RTX 3080 Ti12 GB · best: Qwen3 14B
- RTX 309024 GB · best: Qwen3 32B
- RTX 3090 Ti24 GB · best: Qwen3 32B
- RTX 40608 GB · best: Qwen3 8B
- RTX 4060 Ti8 / 16 GB · best: Qwen3 8B
- RTX 407012 GB · best: Qwen3 14B
- RTX 4070 Super12 GB · best: Qwen3 14B
- RTX 4070 Ti12 GB · best: Qwen3 14B
- RTX 4070 Ti Super16 GB · best: gpt-oss 20B
- RTX 408016 GB · best: gpt-oss 20B
- RTX 4080 Super16 GB · best: gpt-oss 20B
- RTX 409024 GB · best: Qwen3 32B
- RTX 50608 GB · best: Qwen3 8B
- RTX 5060 Ti16 / 8 GB · best: gpt-oss 20B
- RTX 507012 GB · best: Qwen3 14B
- RTX 5070 Ti16 GB · best: gpt-oss 20B
- RTX 508016 GB · best: gpt-oss 20B
- RTX 509032 GB · best: Qwen3 32B
- RTX 3050 Laptop4 / 6 GB · best: Qwen3 4B
- RTX 3060 Laptop6 GB · best: Qwen3 4B
- RTX 3070 Laptop8 GB · best: Qwen3 8B
- RTX 3070 Ti Laptop8 GB · best: Qwen3 8B
- RTX 3080 Laptop8 / 16 GB · best: Qwen3 8B
- RTX 3080 Ti Laptop16 GB · best: gpt-oss 20B
- RTX 4050 Laptop6 GB · best: Qwen3 4B
- RTX 4060 Laptop8 GB · best: Qwen3 8B
- RTX 4070 Laptop8 GB · best: Qwen3 8B
- RTX 4080 Laptop12 GB · best: Qwen3 14B
- RTX 4090 Laptop16 GB · best: gpt-oss 20B
- RTX 5070 Laptop8 GB · best: Qwen3 8B
- RTX 5070 Ti Laptop12 GB · best: Qwen3 14B
- RTX 5080 Laptop16 GB · best: gpt-oss 20B
- RTX 5090 Laptop24 GB · best: Qwen3 32B
- Titan V12 GB · best: Qwen3 14B
- Titan RTX24 GB · best: Qwen3 30B-A3B
- Quadro RTX 40008 GB · best: Qwen3 8B
- Quadro RTX 500016 GB · best: gpt-oss 20B
- Quadro RTX 600024 GB · best: Qwen3 30B-A3B
- Quadro RTX 800048 GB · best: Qwen3 30B-A3B
- RTX A20006 / 12 GB · best: Qwen3 4B
- RTX A400016 GB · best: gpt-oss 20B
- RTX A450020 GB · best: Gemma 3 27B
- RTX A500024 GB · best: Qwen3 32B
- RTX A550024 GB · best: Qwen3 32B
- RTX A600048 GB · best: Qwen3 32B
- RTX 2000 Ada16 GB · best: gpt-oss 20B
- RTX 4000 Ada20 GB · best: gpt-oss 20B
- RTX 4500 Ada24 GB · best: Qwen3 30B-A3B
- RTX 5000 Ada32 GB · best: Qwen3 30B-A3B
- RTX 6000 Ada48 GB · best: Qwen3 32B
- Tesla P4024 GB · best: Qwen3 30B-A3B
- Tesla P10016 / 12 GB · best: gpt-oss 20B
- Tesla T416 GB · best: gpt-oss 20B
- Tesla V10016 / 32 GB · best: gpt-oss 20B
- A1024 GB · best: Qwen3 30B-A3B
- A3024 GB · best: Qwen3 32B
- A4048 GB · best: Qwen3 30B-A3B
- A10040 / 80 GB · best: Qwen3 32B
- H10080 GB · best: gpt-oss 120B
- H200141 GB · best: gpt-oss 120B
- L424 GB · best: Qwen3 30B-A3B
- L40S48 GB · best: Qwen3 32B
- Tesla M4024 / 12 GB · best: Qwen3 30B-A3B
- RTX Pro 4000 Blackwell24 GB · best: Qwen3 30B-A3B
- RTX Pro 4500 Blackwell32 GB · best: Qwen3 32B
- RTX Pro 5000 Blackwell48 GB · best: Qwen3 32B
- RTX Pro 6000 Blackwell96 GB · best: gpt-oss 120B
AMD
- RX 66008 GB · best: Qwen3 8B
- RX 6600 XT8 GB · best: Qwen3 8B
- RX 6700 XT12 GB · best: Qwen3 14B
- RX 680016 GB · best: gpt-oss 20B
- RX 6800 XT16 GB · best: gpt-oss 20B
- RX 6900 XT16 GB · best: gpt-oss 20B
- RX 6950 XT16 GB · best: gpt-oss 20B
- RX 76008 GB · best: Qwen3 8B
- RX 7600 XT16 GB · best: gpt-oss 20B
- RX 7700 XT12 GB · best: Qwen3 14B
- RX 7800 XT16 GB · best: gpt-oss 20B
- RX 7900 GRE16 GB · best: gpt-oss 20B
- RX 7900 XT20 GB · best: Gemma 3 27B
- RX 7900 XTX24 GB · best: Qwen3 32B
- RX 9060 XT16 / 8 GB · best: gpt-oss 20B
- RX 907016 GB · best: gpt-oss 20B
- RX 9070 XT16 GB · best: gpt-oss 20B
- Radeon AI Pro R970032 GB · best: Qwen3 30B-A3B
- Instinct MI5032 / 16 GB · best: Qwen3 32B
- Instinct MI10032 GB · best: Qwen3 32B
- Radeon Pro W680032 GB · best: Qwen3 30B-A3B
- Radeon Pro W780032 GB · best: Qwen3 30B-A3B
- Radeon Pro W790048 GB · best: Qwen3 32B
Intel Arc
Apple
- Apple M116 / 8 GB · best: Qwen3 14B
- Apple M1 Pro16 / 32 GB · best: Qwen3 8B
- Apple M1 Max32 / 64 GB · best: Qwen3 30B-A3B
- Apple M1 Ultra64 / 128 GB · best: Qwen3 32B
- Apple M216 / 8 / 24 GB · best: Qwen3 4B
- Apple M2 Pro16 / 32 GB · best: Qwen3 8B
- Apple M2 Max32 / 64 / 96 GB · best: Qwen3 30B-A3B
- Apple M2 Ultra64 / 128 / 192 GB · best: Qwen3 32B
- Apple M316 / 8 / 24 GB · best: Qwen3 4B
- Apple M3 Pro18 / 36 GB · best: Qwen3 4B
- Apple M3 Max36 / 48 / 64 / 96 / 128 GB · best: Qwen3 30B-A3B
- Apple M3 Ultra96 / 256 / 512 GB · best: gpt-oss 120B
- Apple M416 / 24 / 32 GB · best: Qwen3 4B
- Apple M4 Pro24 / 48 / 64 GB · best: gpt-oss 20B
- Apple M4 Max36 / 48 / 64 / 128 GB · best: Qwen3 30B-A3B
- Apple M516 / 24 / 32 GB · best: Qwen3 4B
- Apple M5 Pro24 / 48 / 64 GB · best: gpt-oss 20B
- Apple M5 Max36 / 48 / 64 / 128 GB · best: Qwen3 30B-A3B