Does Qwen3.8 27B UD-IQ4_XS fit in the NVIDIA RTX 5060 Ti's memory?

Estimate: the UD-IQ4_XS file is 14.3 GB and needs about 15.7 GB with an ~8K context; the NVIDIA RTX 5060 Ti has about 15.5 GB usable, so it does not fully fit (part of it runs on the CPU).

Qwen3.8 27B › NVIDIA RTX 5060 Ti

Qwen3.8 27B on the NVIDIA RTX 5060 Ti

Qwen3.8 27B runs at about 27.6 tokens/s on the NVIDIA RTX 5060 Ti (measured, UD-IQ4_XS, CUDA, single-stream llama-server decode).

Best measured: 27.6 tokens/s generation (UD-IQ4_XS, CUDA). The source gives no prompt-processing figure.

For comparison, our size-scaling estimate for this combination is 26.0 tok/s.

Every run

Quant	Backend	Generate tok/s	Prompt tok/s	Test	Build / by / notes	Source
UD-IQ4_XS	CUDA	27.6	–	decode, llama-server, MTP off	SoundByGAIA 30-token prompt, 8192 context, full GPU offload; mean of two timed requests.	llama.cpp discussion #30328

tg128 = generating 128 tokens; pp512 = reading a 512-token prompt (llama-bench). LocalScore averages several prompt and output lengths, so its numbers run lower than tg128. "decode, llama-server" = single-stream generation speed measured against a running llama-server (no speculative decoding).

Check prices

Amazon Newegg

As an Amazon Associate I earn from qualifying purchases.

No card yet?

Try it on a rented GPU (RunPod)

Affiliate link: if you sign up, RigBench may earn a referral credit. It never affects rankings or numbers.

All results for the NVIDIA RTX 5060 Ti · Qwen3.8 27B on other hardware

Embed this result

HTML

<a href="https://rigbench.pages.dev/model/qwen3-8-27b/on/nvidia-rtx-5060-ti/"><img src="https://rigbench.pages.dev/badge/qwen3-8-27b--nvidia-rtx-5060-ti.svg" alt="Qwen3.8 27B · NVIDIA RTX 5060 Ti: 27.6 tok/s (UD-IQ4_XS, measured)"></a>

Markdown (GitHub README, forums)

[![Qwen3.8 27B · NVIDIA RTX 5060 Ti: 27.6 tok/s (UD-IQ4_XS, measured)](https://rigbench.pages.dev/badge/qwen3-8-27b--nvidia-rtx-5060-ti.svg)](https://rigbench.pages.dev/model/qwen3-8-27b/on/nvidia-rtx-5060-ti/)