Intel · Battlemage (Xe2) · 32 GB · Qwen3.8 27B · Q4_K_M: ~11 tok/s
Intel Arc Pro B70 32GB
32 GB of ECC memory on a new card with a warranty, from Intel's first big Battlemage chip. The hardware is ready and the software is catching up: llama.cpp runs it through SYCL or Vulkan at roughly half the speed its bandwidth suggests, so buy it for capacity, not speed.
Lowest price right now
$1,398
eBay · free shipping · $43 per GB · observed 2026-10-18 02:55 UTC
Buy the cheapest on eBay, $1,398
free shipping · Open box · jc.little.faith.store 7,190 · 100.0%
Price alert
One email per move, item plus shipping. Clear the price to hear about any drop. Manage alerts.
Price history
Tracking started today. The chart appears once there is a second day of data.
Where to buy
| Vendor | Price | Condition | Observed | |
|---|---|---|---|---|
| eBay cheapest active listing |
$1,398 free shipping |
new and used | 2026-10-18 02:55 UTC | Go to eBay |
On eBay now
26 listings passed our filters at the last check. Cheapest delivered first. Seller reputation is feedback count and positive rate.
-
$1,398free shipping -
$1,422.12$1,399 + $23.12 shipping -
$1,515.90$1,500 + $15.90 shipping -
$1,550.49$1,499 + $51.49 shipping -
$1,559.13free shipping -
$1,739.12$1,625.29 + $113.83 shipping
What it runs
For each model, grouped by vendor with the newest first: the largest quant that fits with 8K of context and the estimated speed for one user, at 8K and at 32K of context. On a card, a model too big for its memory can still run with part of the weights in system RAM; we assume 64 GB, and the speed shown accounts for it. The model you picked on the ladder is highlighted. Estimates from the spec sheet, not measurements: see how we estimate.
| Model | Quant | Needs at 8K | Fit | Max context | Decode at 8K | At 32K | Prefill tok/s |
|---|---|---|---|---|---|---|---|
| DeepSeek | |||||||
| DeepSeek V4.1 Flash 552B-A16B Sep 2026 | Q4_K_M | 338 GB | does not fit | ||||
| DeepSeek V4 Flash 284B-A13B Jul 2026 | Q4_K_M | 175 GB | does not fit | ||||
| DeepSeek V3.1 671B-A37B Aug 2025 | Q4_K_M | 411 GB | does not fit | ||||
| Gemma | |||||||
| Gemma 4 12B May 2026 | Q8_0 | 14 GB | fits | 128K+ | ~14 | ~14 | ~770 |
| Gemma 4 31B Mar 2026 | Q6_K | 27 GB | fits | 63K | ~7.1 | ~6.5 | ~300 |
| Gemma 4 26B-A4B Mar 2026 | Q8_0 | 28 GB | fits | 128K+ | ~38 | ~34 | ~2,400 |
| Gemma 3 27B Mar 2025 | Q8_0 | 31 GB | fits | 17K | ~6.1 | ~330 | |
| GLM | |||||||
| GLM-5.3 753B-A40B Aug 2026 | Q4_K_M | 462 GB | does not fit | ||||
| GLM-5.3 Flash 320B-A18B Aug 2026 | Q4_K_M | 197 GB | does not fit | ||||
| GLM-4.6 355B-A32B Sep 2025 | Q4_K_M | 221 GB | does not fit | ||||
| GLM-4.5 Air 106B-A12B Jul 2025 | Q4_K_M | 68 GB | runs with 36 GB in system RAM | ~7.7 | ~6.1 | ~380 | |
| gpt-oss | |||||||
| gpt-oss-120b Aug 2025 | Q4_K_M | 73 GB | runs with 42 GB in system RAM | ~17 | ~14 | ~900 | |
| gpt-oss-20b Aug 2025 | Q8_0 | 24 GB | fits | 128K+ | ~38 | ~30 | ~2,500 |
| Llama | |||||||
| Llama 4 Scout 109B-A17B Apr 2025 | Q4_K_M | 70 GB | runs with 38 GB in system RAM | ~5.5 | ~4.5 | ~270 | |
| Llama 3.3 70B Dec 2024 | IQ2_M | 28 GB | fits | 19K | ~6.8 | ~130 | |
| Llama 3.1 8B Jul 2024 | Q8_0 | 11 GB | fits | 128K+ | ~19 | ~14 | ~1,100 |
| MiMo | |||||||
| MiMo V2.6 Pro 1.02T-A42B Sep 2026 | Q4_K_M | 624 GB | does not fit | ||||
| MiMo V2.6 Flash 309B-A15B Sep 2026 | Q4_K_M | 190 GB | does not fit | ||||
| Mistral | |||||||
| Mistral Medium 3.5 128B Mar 2026 | Q4_K_M | 83 GB | runs with 51 GB in system RAM | ~0.7 | ~36 | ||
| Mistral Small 4 119B-A6.5B Mar 2026 | Q4_K_M | 79 GB | runs with 47 GB in system RAM | ~9.9 | ~700 | ||
| Mistral Small 3.2 24B Jun 2025 | Q8_0 | 28 GB | fits | 30K | ~6.8 | ~380 | |
| Nemotron | |||||||
| Nemotron 3.5 Lightning 30B-A3B Aug 2026 | Q6_K | 27 GB | fits | 128K+ | ~51 | ~49 | ~2,600 |
| Qwen | |||||||
| Qwen3.8 Flash-Next 125B-A6B Aug 2026 | Q4_K_M | 78 GB | runs with 46 GB in system RAM | ~15 | ~14 | ~760 | |
| Qwen3.8 27B Aug 2026 (your pick) | Q8_0 | 31 GB | fits | 28K | ~6.3 | ~340 | |
| Qwen3.6 27B Apr 2026 | Q8_0 | 31 GB | fits | 28K | ~6.3 | ~340 | |
| Qwen3.6 35B-A3B Apr 2026 | Q6_K | 30 GB | fits | 85K | ~56 | ~48 | ~3,100 |
| Qwen3.5 9B Feb 2026 | Q8_0 | 11 GB | fits | 128K+ | ~19 | ~17 | ~1,000 |
| Qwen3 Coder 480B-A35B Jul 2025 | Q4_K_M | 296 GB | does not fit | ||||
| Qwen3 235B-A22B Jul 2025 | IQ2_M | 83 GB | runs with 51 GB in system RAM | ~6.9 | ~5.6 | ~210 | |
| Qwen3 32B Apr 2025 | Q6_K | 31 GB | fits | 13K | ~6.3 | ~280 | |
Specifications
- Memory
- 32 GB GDDR6 ECC
- Memory bandwidth
- 608 GB/s
- Bus width
- 256-bit
- Shader cores
- 4,096
- FP32
- 22.9 TFLOPS
- FP16 (dense)
- 183 TFLOPS
- Power
- 230 W
- Form factor
- Dual slot, full height
- Interface
- PCIe 5.0 x16
- Architecture
- Battlemage (Xe2)
- Launched
- 2026
- MSRP
- $949
Source: manufacturer specifications.