Fastest
unsloth/Qwen3.6-35B-A3B-GGUF:UD-Q4_K_M
35B
Highest token generation speed at 330.1 tok/s.
Hardware Benchmarks
Compare Apple M5 Ultra benchmark results and see which models actually earn the best completed quality scores.
GPU-specific specifications for local LLM planning.
Models we recommend for Apple M5 Ultra.
Fastest
35B
Highest token generation speed at 330.1 tok/s.
Best quality
125B
Highest completed quality score at 91.5.
Largest context tested
Unknown
Largest context window tested on this hardware, at 262,144 tokens. oMLX uses it as a prompt limit (no memory is reserved for the full window), so this doesn't show the full window fits.
Every model the community ran on this hardware. Up is better answers, right is faster. Time to finish counts every token the model wrote, thinking included, so a model that talks a lot ends up further left.
Medians over 31 fully rated runs of 19 models. Each quantization is its own model.
Which models actually run best on the Apple M5 Ultra, by task, from community benchmark data.
Completed quality-ranked benchmark results for Apple M5 Ultra.