More
Best rig by budgetRent before you buySubmit a buildNewsletterFree datasetAdvertise
Submit build

Builds / RTX 3090 24GB

RTX 3090: Qwen3.8 27B Q4_K_M at 38.6 tok/s (llm-speed, llama.cpp b11803, 4 runs)

Last updated:

Source: llm-speed.com ↗More source: llm-speed.com ↗📏 How measured: reported by the original source; RigBase re-checked the numbers at that source
GPU(s)
RTX 3090 24GB
Amazon · eBay (used) · Newegg
VRAM / unified memory (GB)
24
CPU
not stated
Motherboard
not stated
RAM
not stated
PSU
not stated
Other
not stated
Reported total cost
not stated
Power
not stated

Measured speed (generation tokens/sec)

ModelQuantEnginetok/sPrompt processing (tok/s)Details
Qwen3.8 27BQ4_K_Mllama.cpp build 11803 (a37cc4610)38.6483 @ 154-token prompt
measured TTFT: 0.32 s @ 154-token prompt		context: 8K ctx, chat-short (154-token prompt)📏 Lowest of 4 llm-speed runs on 22 Oct 2026 (38.6-38.8)
Qwen3.8 27BQ4_K_Mllama.cpp build 11803 (a37cc4610)38.11,209 @ 3,224-token prompt
est. TTFT @8k ≈ 6.8 s (calculated from measured prefill)
prefill source: llm-speed.com ↗
measured TTFT: 2.7 s @ 3,224-token prompt		context: chat-long (3,224-token prompt, 1,024 out)📏 Lowest chat-long run (38.1-38.4)

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.

Measured by llm-speed on its own RTX 3090, 22 Oct 2026, with the same llama.cpp build and prompts as its RTX 4090 runs, so the two pages compare like for like. Rest of the system not stated.

Source: https://llm-speed.com/can-run/qwen3-8-27b/rtx-3090
https://llm-speed.com/m/qwen3-8-27b

Blank fields = not stated in the source. We never estimate build data; GPU street prices above are separately sourced and dated.

Embed this result

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

[![Measured on RigBase](https://getrigbase.com/badge/llmspeed-rtx-3090-qwen38-27b-q4km-oct22.svg)](https://getrigbase.com/builds/llmspeed-rtx-3090-qwen38-27b-q4km-oct22/)

HTML

<a href="https://getrigbase.com/builds/llmspeed-rtx-3090-qwen38-27b-q4km-oct22/" rel="noopener"><img src="https://getrigbase.com/badge/llmspeed-rtx-3090-qwen38-27b-q4km-oct22.svg" alt="Measured on RigBase" height="20"/></a>

Preview: Measured on RigBase · How to use

Rent before you buy: try this model on a cloud GPU for a few dollars first: Runpod (referral link; sign up with Google to get the new-user bonus)

Quick answers: Can RTX 3090 run Qwen3.8 27B?