Hardware › Strix Halo vs Mac Ultra: measured LLM tok/s on 3 shared models
Strix Halo vs Mac Ultra: measured LLM tok/s on 3 shared models
No shared model has a run on the same engine on both Strix Halo and Mac Ultra: all 3 rows pair different engines. Same model at the same bit-width, best known number per box with source links; not a controlled test.
Snapshot 2026-10-29 · 413 configs on AI boxes and 288 on GPU cards · 121 models. Every number is generated from the tracker snapshot and linked to where it was measured; nothing here is typed by hand.
Strix Halo vs Mac Ultra, measured on 3 shared model and bit-width pairs. No shared model has a run on the same engine on both Strix Halo and Mac Ultra: all 3 rows pair different engines. Median ratio Strix Halo to Mac Ultra, best known number per box: 0.3x. Snapshot 2026-10-29.
Same model, same bit-width, both boxes
| Model | Bits | Strix Halo | Strix Halo source | Mac Ultra | Mac Ultra source | Ratio Strix Halo to Mac Ultra |
|---|---|---|---|---|---|---|
| Qwen3.5-122B-A10B | 4-bit | 21.6 tok/s (UD-Q4_K_XL, llama.cpp/rocm-7.2.4) | source | 65.3 tok/s (MLX-4BIT, mlx) | source | 0.3x |
| Qwen3.6-35B-A3B | 4-bit | 81.3 tok/s (Q4_0, llama.cpp/vulkan) | source | 109.1 tok/s (MLX-4BIT, mlx) | source | 0.7x |
| Qwen3.8-27B | 4-bit | 14.1 tok/s (IQ4_XS, llama.cpp) | source | 42.6 tok/s (MLX-4BIT, mlx) | source | 0.3x |
How to read this
Each cell is the best known single-stream number for that box, taken from its source. It is not a controlled test: backends, quant formats, context length and the person measuring differ between the two sides, and the source link on each cell shows which. A group counts only when both boxes have a plain-decoding run (no speculative decoding) of the same model family at the same bit-width. Models under 7B parameters are left out.
Price in Singapore
- Strix Halo: from S$4,329 today
- Mac Ultra: from S$7,999 today
Questions and answers
- Is Strix Halo faster than Mac Ultra for local LLMs?
- It depends on the model and the engine. No shared model has a run on the same engine on both Strix Halo and Mac Ultra: all 3 rows pair different engines. Across all 3 shared model and bit-width pairs, the median ratio Strix Halo to Mac Ultra, best known number per box, is 0.3x.
- Which model shows the biggest gap between Strix Halo and Mac Ultra?
- Qwen3.5-122B-A10B at 4-bit: Mac Ultra 65.3 tok/s against Strix Halo 21.6 tok/s, a ratio of 3x.
- Is this a fair benchmark of Strix Halo against Mac Ultra?
- No. These are the best known numbers per box, taken from different sources. Backends, quant formats, context length and who measured it all differ, and each row links to its source. Speculative-decoding runs and models under 7B parameters are left out.
- How much do Strix Halo and Mac Ultra cost in Singapore?
- Strix Halo from S$4,329 and Mac Ultra from S$7,999, the lowest listed price across the sellers TokenMark reads daily. The listings are different configurations and memory sizes, so check the where-to-buy list on each hardware page.
Strix Halo hardware page · Mac Ultra hardware page
Other comparisons: Strix Halo vs DGX Spark.
Last updated: snapshot 2026-10-29; newest source check 2026-10-29.
More: all configs · hardware · news · contributors · methodology · llms.txt · llms-full.txt · API (OpenAPI) · JSON snapshot. Built by Altronis, private on-prem AI, Singapore.