Instinct MI100 x4 vs Radeon AI Pro R9700 x3

Best sourced output tok/s per model (single-stream generation) — engine and version, quantization, context and a source link on every row. Hover a value for the quant.

Instinct MI100 x41 sourced run · best 132.4 tok/s

Radeon AI Pro R9700 x39 verified runs · best 223.4 tok/s

Specs and evidence

	Instinct MI100 x4	Radeon AI Pro R9700 x3
VRAM	32 GB	32 GB
Sourced runs on the board	1	9
Models measured	1	9
Best output tok/s	132.4 tok/s	223.4 tok/s
Highest measured peak VRAM	128.0 GB	25.8 GB
Street price (latest snapshot)	—	—

Sourced rows only. Peak VRAM is recorded by the person who ran the benchmark, so it is a floor for what the card can hold, not a specification. Price snapshots.

The measured matrix

Model	Instinct MI100 x4	Radeon AI Pro R9700 x3
gpt-oss-120b	—	223.4 MXFP4
Qwen3.8-27B	132.4 INT8	—
Nemotron-Cascade-2-30B-A3B	—	95.4 Q8_0
Qwen3-VL-8B-Instruct	—	58.8 Q8_0
Muse-Glimmer-30B	—	47.8 UD-Q8_K_XL
Ling-3.0-flash	—	39.3 Q4_0_ROCMFP4_STRIX_LEAN
Ornith-1.5-35B-A3B	—	35.4 BF16
Ornith-1.0-9B	—	31.2 BF16
Qwen3.8-Flash-Next	—	28.1 UD-IQ3_XXS
gemma-4-31B-it	—	23.9 Q8_0

10 rows · 1 models on Instinct MI100 x4, 9 on Radeon AI Pro R9700 x3 — no verified overlap yet, each side shows its own bests · data as of Oct 17, 2026. Full rules: methodology.

Related reading: RTX 5090 vs 3090 · quantization · Value index · Full board