Builds / Intel Arc Pro B60 24GB
Intel Arc Pro B60 24GB + Ryzen 7 3700X: Qwen3.8 27B Q4_K_M at 16 tok/s (14.5 at 32K), 42.3 with DFlash2
Last updated:
- GPU(s)
- Intel Arc Pro B60 24GB
Amazon · eBay (used) · Newegg - VRAM / unified memory (GB)
- 24
- CPU
- AMD Ryzen 7 3700X
Amazon · eBay (used) · Newegg - Motherboard
- B550
Amazon · eBay (used) · Newegg - RAM
- 32 GB DDR4
Amazon · eBay (used) · Newegg - PSU
- not stated
- Other
- Windows 11, Intel driver 32.0.101.8976; llama.cpp Vulkan
- Reported total cost
- not stated
- Power
- not stated
Measured speed (generation tokens/sec)
| Model | Quant | Engine | tok/s | Prompt processing (tok/s) | Details |
|---|---|---|---|---|---|
| Qwen3.8 27B | UD-Q4_K_M, KV f16, no drafter | llama.cpp Vulkan b11524 (86a283532), stock | 16.01 | 589 @ 2,048-token prompt est. TTFT @8k ≈ 13.9 s (calculated from measured prefill) | context: empty context, llama-bench tg32📏 llama-bench tg32, single run (-r 1) |
| Qwen3.8 27B | UD-Q4_K_M, KV f16, no drafter | llama.cpp Vulkan b11524 (86a283532), stock | 14.45 | 118 @ 2,048 @32K depth-token prompt est. TTFT @8k ≈ 69.4 s (calculated from measured prefill) | context: 32K context depth, llama-bench tg32📏 Same llama-bench row at 32K depth (sustained figure); 15.14 at 16K |
| Qwen3.8 27B | UD-Q4_K_M + DFlash2 draft model (n=7), KV q8 draft | llama.cpp Vulkan + Intel flash-attention prefill kernel (arc-llm-lab build) | 42.3 | not stated | ⚡ spec-decode (DFlash)context: 64K KV window allocated, 8 eval requests📏 Eval-suite throughput: generated tokens / total wall-clock incl. prompt processing (the source's conservative headline; per-request decode 47-58, mean 53.3) |
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.
Elie Dagher, arc-llm-lab (Oct 2026). Single runs on one workstation, raw logs in the repo. The source saw decode fall from 52.7 to 22.2 tok/s while an idle Ollama model was still resident on the card, so keep the GPU free.
https://github.com/edagher92-coder/arc-llm-lab
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/arc-pro-b60-24gb-ryzen-3700x-qwen38-27b-vulkan/)
HTML
<a href="https://getrigbase.com/builds/arc-pro-b60-24gb-ryzen-3700x-qwen38-27b-vulkan/" rel="noopener"><img src="https://getrigbase.com/badge/arc-pro-b60-24gb-ryzen-3700x-qwen38-27b-vulkan.svg" alt="Measured on RigBase" height="20"/></a>
Preview: 
· How to use