WHAT THIS CARD IS WORTH
Radeon Instinct MI100 holds 337 of the models in our catalogue and is, in practice, a Q4_K_M card — the largest it takes is Phi-3.5 MoE 42B at Q4_K_M.
TOKENS/SEC PER $100
2.6
8B at Q4_K_M, so cards compare like for like
VRAM PER $100
0.64 GB
what memory costs on this card
THE BIGGEST IT TAKES
AGAINST CARDS OF SIMILAR MEMORY
Radeon Instinct MI10032 GB$5,0002.6 tok/s per $100
Neighbours in memory rather than in price: memory decides whether a card can do the job at all, so two cards of the same size at different prices is the comparison you are making. Ordered by memory, then bandwidth — not by the value column, which is worked out from bandwidth and price alone and therefore rewards a cheap card whatever its software stack does to that bandwidth in practice. Read it as one input, not as a ranking.
▸ DEVICE UNDER TEST
AMD Radeon Instinct MI100 — 32 GB VRAM.
▸ RADEON INSTINCT MI100 SPEC
- BRAND
- AMD
- VRAM
- 32 GB HBM2
- BANDWIDTH
- 1230 GB/s
- FP16 COMPUTE
- 184.6 TFLOPS
- FP32 COMPUTE
- 23.1 TFLOPS
- STREAM PROCESSORS
- 7,680
- TDP
- 300 W
- ARCHITECTURE
- CDNA 1.0
- MSRP
- $5000
▸ AI CAPABILITY
337/ 449 models @ Q4
With 32 GB VRAM and 1230 GB/s bandwidth, this GPU handles models up to 41.9B parameters.
Speed ≈ bandwidth / model_size × efficiency. A 7B model at Q4 runs at ~141 tok/s.
RENT IT — LIVE PRICES
Checking what the clouds are charging…
§ 01TOP MODELS FOR RADEON INSTINCT MI100
337 FIT · SHOWING 20| MODEL | SIZE | VRAM Q4 | TOK/S | AVG |
|---|---|---|---|---|
| Phi-3.5 MoE 42B | 41.9B | 26.1 GB | 166 | 56.7 |
| Falcon 40B | 40B | 24.9 GB | 27 | 20.9 |
| InternVL3 38B | 38B | 23.7 GB | 29 | 78.9 |
| Seed-OSS 36B Instruct | 36B | 22.5 GB | 30 | 54.4 |
| c4ai-command-r-v01 35B | 35B | 21.9 GB | 31 | 27.5 |
| Qwen 3.5 35B A3B | 35B | 21.9 GB | 364 | 53.3 |
| Qwen 3.6 35B A3B | 35B | 21.9 GB | 364 | 53.9 |
| Nous Capybara 34B | 34.4B | 21.5 GB | 32 | 42.0 |
| Yi-1.5 34B | 34.4B | 21.5 GB | 32 | 45.3 |
| Falcon-H1 34B | 34B | 21.3 GB | 32 | 66.1 |
| CodeLlama 34B | 34B | 21.3 GB | 32 | 25.4 |
| Nous Hermes 2 34B | 34B | 21.3 GB | 32 | 47.0 |
| Phind CodeLlama 34B | 34B | 21.3 GB | 32 | 68.1 |
| LLaVA-1.6 Yi 34B | 34B | 21.3 GB | 32 | 47.4 |
| WizardCoder Python 34B | 34B | 21.3 GB | 32 | 73.2 |
| Yi 34B | 34B | 21.3 GB | 32 | 33.4 |
| Qwen3-VL 32B Instruct | 33.36B | 20.9 GB | 33 | 44.6 |
| DeepSeek Coder 33B | 33B | 20.7 GB | 33 | 26.0 |
| Vicuna 33B | 33B | 20.7 GB | 33 | 17.2 |
| LLaMA 1 30B | 33B | 20.7 GB | 33 | 17.8 |