Can the GeForce RTX 5070 Ti run gpt-oss-20b?
Yes. The Q4_K_M GGUF of gpt-oss-20b is 11.67 GB (source). With our 1.5 GB allowance for runtime and short-context cache it needs about 13.2 GB, leaving 2.8 GB of the card's 16 GB free. Expected speed: ~339 tok/s estimate. Lowest tracked price: low $1,120, avg $1,792 (new) source, 2026-10-16, which works out to $3 per tok/s (#2 of 12 priced GPUs that fit this model).
Last updated Oct 16, 2026. Speed is a bandwidth-based estimate (896 GB/s, TechPowerUp), not a benchmark. Long contexts need more memory than our allowance.
Tight fit: under 3 GB spare means long prompts or a bigger context window may not fit. Consider a smaller quant or a card with more VRAM.
Next size up: gemma-3-27b-it (16.55 GB at Q4_K_M) needs ~18.1 GB and does not fit on one NVIDIA GeForce RTX 5070 Ti. See what runs gemma-3-27b-it.
Every priced GPU that runs gpt-oss-20b, by cost per token/sec
| GPU | VRAM | Speed estimate | Low price | $ per tok/s |
|---|---|---|---|---|
| AMD Radeon RX 9070 | 16 GB | ~244 tok/s | $570 (new) | $2 |
| NVIDIA GeForce RTX 5070 Ti | 16 GB | ~339 tok/s | $1,120 (new) | $3 |
| AMD Radeon RX 9070 XT | 16 GB | ~244 tok/s | $810 (new) | $3 |
| AMD Radeon RX 7800 XT | 16 GB | ~236 tok/s | $599 (new) | $3 |
| NVIDIA GeForce RTX 5080 | 16 GB | ~363 tok/s | $1,600 (new) | $4 |
| AMD Radeon RX 7900 XTX | 24 GB | ~363 tok/s | $1,350 (new) | $4 |
| NVIDIA GeForce RTX 5060 Ti 16 GB | 16 GB | ~169 tok/s | $780 (new) | $5 |
| NVIDIA GeForce RTX 3090 | 24 GB | ~354 tok/s | $1,889 (used) | $5 |
| AMD Radeon RX 9060 XT 16 GB | 16 GB | ~122 tok/s | $620 (new) | $5 |
| NVIDIA GeForce RTX 4090 | 24 GB | ~382 tok/s | $3,200 (used) | $8 |
| NVIDIA GeForce RTX 5090 | 32 GB | ~677 tok/s | $6,500 (new) | $10 |
| NVIDIA RTX A6000 | 48 GB | ~291 tok/s | $6,170 (new) | $21 |
Prices: picksbycard read of Newegg listings, 2026-10-16. Prices move daily.
Buy on Amazon As an Amazon Associate I earn from qualifying purchases.