Single Intel Arc Pro B70 32 GB (neural.download lab)
| GPU(s) | Intel Arc Pro B70 32 GB Amazon · eBay (used) · Newegg |
|---|---|
| VRAM / unified memory (GB) | 32 |
| CPU | |
| Motherboard | |
| RAM | |
| PSU | |
| Other | One card of the lab's four-B70 Linux host; llama.cpp SYCL / oneAPI |
| Reported total cost | |
| Power |
Measured speed (generation tokens/sec)
📏 How measured: reported by the original source; RigBase re-checked the numbers at that source
| Model | Quant | Engine | tok/s | Details |
|---|---|---|---|---|
| Gemma 4 26B-A4B | UD-Q8_K_XL | llama.cpp SYCL (pinned build, oneAPI 2026.0) | 122.16 | 📏 Single-user, cache-cold decode median (lab package/repro README) |
| Ornith 1.5 35B-A3B | Q4_K_M | llama.cpp SYCL + neural.download fusion patches; author marks this headline 'pending' | 131.46 | ⏳ author marks pending📏 Single-user, cache-cold decode median (lab package/repro README) |
| Qwen3.8 27B | Q4_K_M | llama.cpp SYCL + neural.download TP1 stack, no speculation | 27.8 | 📏 Single-user, cache-cold decode median (lab package/repro README) |
Single-user, cache-cold decode medians from the lab's package/repro READMEs. Ornith 35B-A3B headline is marked 'pending' on neural.download (performance measured, install checks not finished). Host CPU/RAM not recorded in the package docs. Found via localmaxxing (discovery only); numbers re-checked at the original source on 2026-10-24.
Source: https://github.com/steveseguin/b70-optimization-lab
https://neural.download/models/qwen38-27b-q4km-tp1-b70.html
https://neural.download/models/ornith-15-35b-a3b-q4km-b70.html
Blank fields = not stated in the source. We never estimate build data; GPU street prices above are separately sourced and dated.
Quick answers: Can Intel Arc Pro B70 run Gemma 4 26B-A4B? · Can Intel Arc Pro B70 run Ornith 1.5 35B-A3B? · Can Intel Arc Pro B70 run Qwen3.8 27B?