Can OLMoE 1B 7B 0924 run on the RTX 5070 Ti?
Yes. At 8K context the best fit is Q8_0, using about 8.4 GB of the 15.2 GB usable on the RTX 5070 Ti. Estimated speed ceiling: ~657 tokens/sec.
Memory by quant and context
| Quant | 4K ctx | Max tok/s (est., 8K) |
|---|---|---|
| 2-bit (Q2_K) • | 3.4 GB Fits | 1887 |
| 3-bit (Q3_K_M) • | 4.1 GB Fits | 1446 |
| 4-bit (Q4_K_M) • | 4.9 GB Fits | 1148 |
| 5-bit (Q5_K_M) • | 5.6 GB Fits | 982 |
| 6-bit (Q6_K) • | 6.3 GB Fits | 851 |
| 8-bit (Q8_0) • | 7.9 GB Fits | 657 |
| 16-bit (F16/BF16) • | 13.9 GB Tight | 349 |
• real GGUF file size from allenai/OLMoE-1B-7B-0924-GGUF. Speed is a bandwidth ceiling (896 GB/s), not a benchmark. Methodology.
More
OLMoE 1B 7B 0924 on every GPU and Mac · Every model that runs on the RTX 5070 Ti · Other models that fit 16 GB VRAM · Try other settings in the calculator
Data: Hugging Face Hub (allenai/OLMoE-1B-7B-0924), NVIDIA specs. Data as of 2026-11-17.