▸ DEVICE UNDER TEST
AMD Radeon Instinct MI250X — 128 GB VRAM.
▸ RADEON INSTINCT MI250X SPEC
- BRAND
- AMD
- VRAM
- 128 GB HBM2e
- BANDWIDTH
- 3280 GB/s
- FP16 COMPUTE
- 383 TFLOPS
- FP32 COMPUTE
- 47.9 TFLOPS
- STREAM PROCESSORS
- 14,080
- TDP
- 500 W
- ARCHITECTURE
- CDNA 2.0
- MSRP
- $15000
▸ AI CAPABILITY
302/ 331 models @ Q4
With 128 GB VRAM and 3280 GB/s bandwidth, this GPU handles models up to 180B parameters.
Speed ≈ bandwidth / model_size × efficiency. A 7B model at Q4 runs at ~375 tok/s.
§ 01TOP MODELS FOR RADEON INSTINCT MI250X
302 FIT · SHOWING 20| MODEL | SIZE | VRAM Q4 | TOK/S | AVG |
|---|---|---|---|---|
| Falcon 180B | 180B | 110.5 GB | 16 | 51.3 |
| bloom 176.2B | 176.2B | 108.2 GB | 17 | 15.0 |
| dots.llm1.inst 142.8B | 142.8B | 87.8 GB | 20 | — |
| WizardLM 2 8x22B | 141B | 86.7 GB | 75 | 42.4 |
| Mixtral-8x22B | 140.6B | 86.4 GB | 75 | 31.9 |
| DBRX 132B | 132B | 81.2 GB | 81 | 46.3 |
| Qwen3.5-122B-A10B | 125.1B | 77.0 GB | 23 | 45.5 |
| Pixtral Large 124B | 124B | 76.3 GB | 24 | 39.3 |
| Mistral-Large 123B | 123B | 75.7 GB | 24 | 33.5 |
| Devstral 2 123B | 123B | 75.7 GB | 24 | 38.1 |
| Qwen 3.5 122B A10B | 122B | 75.1 GB | 292 | 56.8 |
| Nemotron 3 Super 120B | 120B | 73.8 GB | 243 | 57.3 |
| Nemotron 3 Super 120B-A12B | 120B | 73.8 GB | 243 | 53.2 |
| Mistral Small 4 119B | 119B | 73.2 GB | 449 | 50.2 |
| GPT-OSS 120B | 117B | 72.0 GB | 572 | 54.1 |
| Command A 111B | 111B | 68.3 GB | 26 | 27.6 |
| GLM 4.5 Air | 110B | 67.7 GB | 243 | 51.0 |
| Qwen 1.5 110B | 110B | 67.7 GB | 27 | 33.4 |
| Llama 4 Scout 17B-16E | 109B | 67.1 GB | 172 | 33.9 |
| Sarvam 105B | 105B | 64.7 GB | 28 | 48.0 |