M3 Ultra (80-core GPU, 256 GB) — LLM Benchmarks

Measured LLM inference benchmarks for M3 Ultra (80-core GPU, 256 GB). Tokens per second across 3 models and multiple quantizations. Real runs, not estimates.

3 Benchmark rows

3 Models tested

177.9 Fastest avg tok/s (Llama 3.2 1B Instruct)

0 Factory-lab verified rows

Benchmark results for M3 Ultra (80-core GPU, 256 GB)

Rows sorted by avg tok/s descending. Click source badge to see original measurement page.

Model	Quant	RAM req.	Context	Avg tok/s	Prompt tok/s	Runtime	Source
Llama 3.2 1B Instruct	Q4_K - Medium	—	—	177.9 tok/s	4999.5 tok/s	—	ref
Llama 3.1 8B Instruct	Q4_K - Medium	—	—	63.3 tok/s	1062.2 tok/s	—	ref
Qwen 2.5 14B Instruct	Q4_K - Medium	—	—	36.7 tok/s	568.3 tok/s	—	ref

Models tested on this chip

Llama 3.1 8B Instruct Llama 3.2 1B Instruct Qwen 2.5 14B Instruct

Data

benchmarks.json — full dataset · chips.json — chip summaries · benchmarks.csv — CSV export

Want to contribute a benchmark for this chip? Data sourced from factory lab measurements and community reference runs. See all chips →