Builds / RTX 5070 Ti
RTX 5070 Ti 16 GB (Core Ultra 9 285K, llama.cpp b12128 CUDA): gpt-oss 20B at 236.7 tok/s, Gemma 4 E4B Q4_K_M 178.6, Qwen3.5 9B Q6_K 107.5
Last updated:
- GPU(s)
- RTX 5070 Ti
Amazon · eBay (used) · Newegg - VRAM / unified memory (GB)
- 16
- CPU
- Intel Core Ultra 9 285K
Amazon · eBay (used) · Newegg - Motherboard
- not stated
- RAM
- not stated
- PSU
- not stated
- Other
- Ubuntu 26.04 LTS, llama.cpp b12128 CUDA (CUDA was the fastest backend on this card in every text test); card sits in a PCIe 5.0 x2 slot (source estimates <3% effect on decode, slower model loading); 8 fastest CPU cores pinned
- Reported total cost
- not stated
- Power
- 1.58 tokens per watt on gpt-oss 20B decode (driver sensor); under 4 W idle
Measured speed (generation tokens/sec)
| Model | Quant | Engine | tok/s | Prompt processing (tok/s) | Details |
|---|---|---|---|---|---|
| gpt-oss 20B | MXFP4 | llama.cpp b12128 (CUDA) | 236.7 | 8,442 @ 4,096-token prompt est. TTFT @8k ≈ 0.97 s (calculated from measured prefill) | context: empty context; 208.7 at 16,384 depth📏 write 128 tokens, 5 repetitions (tg128), model fully in VRAM, load time excluded |
| Qwen3.5 9B | Q6_K | llama.cpp b12128 (CUDA) | 107.5 | 4,987 @ 4,096-token prompt est. TTFT @8k ≈ 1.6 s (calculated from measured prefill) | context: empty context📏 tg128, 5 repetitions (Vulkan crashed at empty context on this card; 96.5 with 4,096 context) |
| Gemma 4 E4B legacy | Q4_K_M | llama.cpp b12128 (CUDA) | 178.6 | 10,724 @ 4,096-token prompt est. TTFT @8k ≈ 0.76 s (calculated from measured prefill) | context: empty context📏 tg128, 5 repetitions |
Prompt processing = prefill speed (how fast the prompt is read) as reported by the source, with the prompt length it was measured at; blank = the source does not report it. est. TTFT @8k is calculated from the measured prefill speed (8,192 ÷ prefill tok/s), shown only when prefill was measured on a prompt of 512+ tokens. It is not a measurement and ignores load time and caching. measured TTFT is the source's own time-to-first-token. Why prefill matters for coding assistants.
Vendor-run: GEARVN AI Lab (GEARVN is a Vietnamese PC retailer that sells all three cards), article published 5-6 Nov 2026, measured 6 Nov 2026. Same llama.cpp b12128, same six GGUF files (sha256 checked on all three machines), 8 fastest CPU cores pinned, other GPU services paused; re-running on the single slowest core changed speed by at most 3.5%, so the gaps are the GPUs. Each card ran its fastest backend. Multi-user, gpt-oss 20B on CUDA (1,024 tokens in, 256 out per user): 1 user 230.6, 4 users 533.7, 8 users 666.7, 16 users 1,255.7 tok/s total (about 78 per user). Same test on the Radeon AI PRO R9700 32 GB (Threadripper PRO 7965WX): gpt-oss 20B 181.3 (Vulkan), Gemma 4 E4B 129.6, Qwen3.5 9B 75.9. Models over 16 GB (Qwen3.6 35B-A3B, Gemma 4 26B-A4B, Qwen3.8 27B) were not run on this card.
- 1× RTX 5070 Ti: ≈€1,245 each (≈$1,395), new, approx. price as of 2026-10-24, lowest listing on billiger.de (PNY GeForce RTX 5070 Ti 16 GB (listed as "GDDR6"))
New retail prices in Germany incl. VAT; used cards are often much cheaper. Prices change daily.
Blank fields = not stated in the source. We never estimate build data; GPU street prices above are separately sourced and dated.
Copy into a README or blog. Links back to this build. See badge docs (CC BY 4.0 attribution). The SVG does not invent speeds — numbers stay on the build page.
Markdown
[](https://getrigbase.com/builds/rtx-5070-ti-16gb-core-ultra-9-285k-gpt-oss-gemma4-e4b-qwen35-gearvn/)
HTML
<a href="https://getrigbase.com/builds/rtx-5070-ti-16gb-core-ultra-9-285k-gpt-oss-gemma4-e4b-qwen35-gearvn/" rel="noopener"><img src="https://getrigbase.com/badge/rtx-5070-ti-16gb-core-ultra-9-285k-gpt-oss-gemma4-e4b-qwen35-gearvn.svg" alt="Measured on RigBase" height="20"/></a>
Preview: 
· How to use
Quick answers: Can RTX 5070 Ti run Qwen3.5 9B? · Can RTX 5070 Ti run gpt-oss 20B?