Compare

Loading community benchmarks…

Community benchmarks

Compare

Put two chips, runtimes, or runtime versions side by side. Everything else is held fixed or called out, so a ratio is worth exactly what the matched runs say.

Add your benchmark

Runs compared: 7
Like-for-like pair: 1
Chips available: 5

Compare

Side A

Side B

Model

Runtime

Runtime version

Quantisation

Backend

Prefill size

Decode size

Cooldown

4 runs vs 3 runs · 1 like-for-like pair

Held fixedModel: Qwen3.5 35B A3BPrefill: PP512Decode: TG128

A side mixes quantisations. Pick a quantisation to compare the same weights.

Chip comparison

NVIDIA GB10 vs Apple M1 Max

Ratios read side A relative to side B. Geometric mean over 1 like-for-like pair.

Decode: 0.74×-25.8%
Prefill: 8.71×+771%

Side A

NVIDIA GB10

4 runs

1 contributor

Latest 19 Sep 2026

Decode · TG128: 69.5tok/s median
best 92.4 · worst 54.2
Prefill · PP512: 2,437tok/s median
best 2,521 · worst 2,093

Models: Qwen3.5-35B-A3B-cuda-q4mix, Qwen3.5-35B-A3B-cuda-q8, Qwen3.5-35B-A3B-Q4
Runtimes: BaseRT
Versions: BaseRT 0.2.6
Quantisations: Q4
Backends: CUDA
Cooldown: off
Conditioning: warmup_only
Decode workload: TG128 @ 1 ctx
Harness schema: basert-benchmark-harness/1

Side B

Apple M1 Max

3 runs

1 contributor

Latest 19 Sep 2026

Decode · TG128: 72.2tok/s median
best 93.7 · worst 72.0
Prefill · PP512: 251tok/s median
best 280 · worst 250

Models: Qwen3.5-35B-A3B-Q4, Qwen3.5-35B-A3B-Q8
Runtimes: BaseRT
Versions: BaseRT 0.2.6
Quantisations: Q4, Q8
Backends: Metal
Cooldown: off
Conditioning: warmup_only
Decode workload: TG128 @ 1 ctx
Harness schema: basert-benchmark-harness/1

What differs besides the chip

Facet	NVIDIA GB10	Apple M1 Max
Model	Qwen3.5-35B-A3B-cuda-q4mix, Qwen3.5-35B-A3B-cuda-q8, Qwen3.5-35B-A3B-Q4	Qwen3.5-35B-A3B-Q4, Qwen3.5-35B-A3B-Q8
Quantisation	Q4	Q4, Q8
Backend	CUDA	Metal

Like-for-like pairs

Each row is a configuration present on both sides. Values are per-cell medians; ratios read NVIDIA GB10 relative to Apple M1 Max.

Configuration	Decode A	Decode B	Ratio	Prefill A	Prefill B	Ratio	Runs A / B
Qwen3.5-35B-A3B BaseRTQ4	69.5	93.7	0.74×	2,437	280	8.71×	4 / 1

Side A · NVIDIA GB10

Top 4 of 4 runs by decode throughput.

Model / format	Chip / backend	Decode	Prefill	Contributor	Date	Report
Qwen3.5-35B-A3B-cuda-q4mix BaseRT 0.2.6Q4	NVIDIA GB10 CUDA	92.4 TG128 @ 1 ctx	2,521 PP512	basecompute	19 Sep 2026	View Benchmark
Qwen3.5-35B-A3B-Q4 BaseRT 0.2.6Q4	NVIDIA GB10 CUDA	84.4 TG128 @ 1 ctx	2,093 PP512	basecompute	19 Sep 2026	View Benchmark
Qwen3.5-35B-A3B-cuda-q8 BaseRT 0.2.6Q4	NVIDIA GB10 CUDA	54.7 TG128 @ 1 ctx	2,447 PP512	basecompute	19 Sep 2026	View Benchmark
Qwen3.5-35B-A3B-cuda-q8 BaseRT 0.2.6Q4	NVIDIA GB10 CUDA	54.2 TG128 @ 1 ctx	2,428 PP512	basecompute	19 Sep 2026	View Benchmark

Side B · Apple M1 Max

Top 3 of 3 runs by decode throughput.

Model / format	Chip / backend	Decode	Prefill	Contributor	Date	Report
Qwen3.5-35B-A3B-Q4 BaseRT 0.2.6Q4	Apple M1 Max Metal	93.7 TG128 @ 1 ctx	280 PP512	basecompute	19 Sep 2026	View Benchmark
Qwen3.5-35B-A3B-Q8 BaseRT 0.2.6Q8	Apple M1 Max Metal	72.2 TG128 @ 1 ctx	251 PP512	basecompute	19 Sep 2026	View Benchmark
Qwen3.5-35B-A3B-Q8 BaseRT 0.2.6Q8	Apple M1 Max Metal	72.0 TG128 @ 1 ctx	250 PP512	basecompute	19 Sep 2026	View Benchmark