← CanIRunLocal · All hardware · All models
Alibaba · 80B parameters (3B active) · chat, reasoning.
Ollama weight files: Q4_K_M 50 GB (qwen3-next:80b); Q8_0 85 GB (qwen3-next:80b-a3b-instruct-q8_0).
Yes on 25 machines in this list. The smallest is M1 Max 64GB: Q4_K_M, about 47.2 GB, roughly 95 tokens per second, via LM Studio. lms get mlx-community/Qwen3-Next-80B-A3B-Instruct-4bit
ollama run qwen3-next:80b-a3b-instruct-q8_0ollama run qwen3-next:80b-a3b-instruct-q8_0ollama run qwen3-next:80b-a3b-instruct-q8_0ollama run qwen3-next:80b-a3b-instruct-q8_0lms get unsloth/Qwen3-Next-80B-A3B-Instruct-GGUF@Q8_0llama-server -hf unsloth/Qwen3-Next-80B-A3B-Instruct-GGUF:Q8_0 -c 8192vllm serve RedHatAI/Qwen3-Next-80B-A3B-Instruct-FP8-dynamic --max-model-len 8192 --gpu-memory-utilization 0.9| Hardware | Runtime | Quant | Memory | Fit | Speed | Command |
|---|---|---|---|---|---|---|
| M1 Max 64GB | LM Studio | MLX 4bit | 47.2 GB | fits in fast memory | ~95 tok/s | lms get mlx-community/Qwen3-Next-80B-A3B-Instruct-4bit |
| M2 Max 64GB | LM Studio | MLX 4bit | 47.2 GB | fits in fast memory | ~95 tok/s | lms get mlx-community/Qwen3-Next-80B-A3B-Instruct-4bit |
| M3 Max 64GB | LM Studio | MLX 4bit | 47.2 GB | fits in fast memory | ~95 tok/s | lms get mlx-community/Qwen3-Next-80B-A3B-Instruct-4bit |
| M4 Max 64GB | LM Studio | MLX 4bit | 47.2 GB | fits in fast memory | ~130 tok/s | lms get mlx-community/Qwen3-Next-80B-A3B-Instruct-4bit |
| Mac Studio M4 Max (40c GPU) 64GB | LM Studio | MLX 4bit | 47.2 GB | fits in fast memory | ~130 tok/s | lms get mlx-community/Qwen3-Next-80B-A3B-Instruct-4bit |
| Mac Studio M5 Max (40c GPU) 64GB | LM Studio | MLX 4bit | 47.2 GB | fits in fast memory | ~146 tok/s | lms get mlx-community/Qwen3-Next-80B-A3B-Instruct-4bit |
| DGX Spark 64GB (ASUS, Dell, HP, MSI, Acer, Gigabyte) | Ollama | Q4_K_M | 52.4 GB | fits in fast memory | ~84 tok/s | ollama run qwen3-next:80b |
| Mac Studio M3 Ultra 96GB | Ollama | Q4_K_M | 52.4 GB | fits in fast memory | ~175 tok/s | ollama run qwen3-next:80b |
| Mac Studio M5 Ultra 96GB | Ollama | Q4_K_M | 52.4 GB | fits in fast memory | ~256 tok/s | ollama run qwen3-next:80b |
| M3 Max 128GB | Ollama | Q8_0 | 87.3 GB | fits in fast memory | ~50 tok/s | ollama run qwen3-next:80b-a3b-instruct-q8_0 |
| M4 Max 128GB | Ollama | Q8_0 | 87.3 GB | fits in fast memory | ~69 tok/s | ollama run qwen3-next:80b-a3b-instruct-q8_0 |
| Mac Studio M4 Max (40c GPU) 128GB | Ollama | Q8_0 | 87.3 GB | fits in fast memory | ~69 tok/s | ollama run qwen3-next:80b-a3b-instruct-q8_0 |
| Mac Studio M5 Max (40c GPU) 128GB | Ollama | Q8_0 | 87.3 GB | fits in fast memory | ~77 tok/s | ollama run qwen3-next:80b-a3b-instruct-q8_0 |
| DGX Spark 128GB (NVIDIA, ASUS, Dell, Lenovo, HP, MSI, Acer, Gigabyte) | Ollama | Q8_0 | 87.3 GB | fits in fast memory | ~49 tok/s | ollama run qwen3-next:80b-a3b-instruct-q8_0 |
| M2 Ultra 192GB | Ollama | Q8_0 | 87.3 GB | fits in fast memory | ~100 tok/s | ollama run qwen3-next:80b-a3b-instruct-q8_0 |
| 2x Mac Studio M5 Max (40c GPU) 128GB | llama.cpp | Q8_0 GGUF | 87.2 GB | fits in fast memory | ~66 tok/s | llama-server -hf unsloth/Qwen3-Next-80B-A3B-Instruct-GGUF:Q8_0 --rpc node2:50052 -c 8192 # run rpc-server on each other node |
| Mac Studio M3 Ultra 256GB | Ollama | Q8_0 | 87.3 GB | fits in fast memory | ~103 tok/s | ollama run qwen3-next:80b-a3b-instruct-q8_0 |
| Mac Studio M5 Ultra (64c GPU) 256GB | Ollama | Q8_0 | 87.3 GB | fits in fast memory | ~151 tok/s | ollama run qwen3-next:80b-a3b-instruct-q8_0 |
| Mac Studio M5 Ultra (80c GPU) 256GB | Ollama | Q8_0 | 87.3 GB | fits in fast memory | ~151 tok/s | ollama run qwen3-next:80b-a3b-instruct-q8_0 |
| 2x DGX Spark 128GB | llama.cpp | Q8_0 GGUF | 87.2 GB | fits in fast memory | ~42 tok/s | llama-server -hf unsloth/Qwen3-Next-80B-A3B-Instruct-GGUF:Q8_0 --rpc node2:50052 -c 8192 # run rpc-server on each other node |
| 4x Mac Studio M5 Max (40c GPU) 128GB | llama.cpp | Q8_0 GGUF | 87.2 GB | fits in fast memory | ~58 tok/s | llama-server -hf unsloth/Qwen3-Next-80B-A3B-Instruct-GGUF:Q8_0 --rpc node2:50052,node3:50052,node4:50052 -c 8192 # run rpc-server on each other node |
| 2x Mac Studio M5 Ultra (80c GPU) 256GB | llama.cpp | Q8_0 GGUF | 87.2 GB | fits in fast memory | ~128 tok/s | llama-server -hf unsloth/Qwen3-Next-80B-A3B-Instruct-GGUF:Q8_0 --rpc node2:50052 -c 8192 # run rpc-server on each other node |
| Mac Studio M3 Ultra (80c GPU) 512GB | Ollama | Q8_0 | 87.3 GB | fits in fast memory | ~103 tok/s | ollama run qwen3-next:80b-a3b-instruct-q8_0 |
| 4x DGX Spark 128GB | llama.cpp | Q8_0 GGUF | 87.2 GB | fits in fast memory | ~37 tok/s | llama-server -hf unsloth/Qwen3-Next-80B-A3B-Instruct-GGUF:Q8_0 --rpc node2:50052,node3:50052,node4:50052 -c 8192 # run rpc-server on each other node |
| 4x Mac Studio M5 Ultra (80c GPU) 256GB | llama.cpp | Q8_0 GGUF | 87.2 GB | fits in fast memory | ~113 tok/s | llama-server -hf unsloth/Qwen3-Next-80B-A3B-Instruct-GGUF:Q8_0 --rpc node2:50052,node3:50052,node4:50052 -c 8192 # run rpc-server on each other node |
Estimates, not benchmarks. Model sizes are ollama.com download sizes; memory needed adds headroom for ~8k tokens of context; speeds are rough figures from memory bandwidth. Data updated 2026-10-16.
A project by OnDevice Labs · llms.txt