Intel Arc B580
12 GB · 456 GB/s · 190 W
In short
The Intel Arc B580 has 12 GB of GDDR6 memory with 456 GB/s of bandwidth, which sets how fast it generates text. At 32k context it runs 47 of the 113 open models we track fully in memory; the most capable is Qwen 3.8 27B at UD-Q5_K_M, at about 10 tok/s (estimated). The best new price for the Intel Arc B580 in Canada is $550 at eBay; used, it starts at $450 (asking price).
Updated . Speeds are estimates; how they are worked out.
Prices
Every retailer we track in Canada, cheapest first. Buy links may earn us a commission.
Price history
Lowest price shown each day over the last 90 days.
Not enough price history yet.
What two of them run
Every model in the catalogue, with the best quant for this hardware. Estimates, not measurements.
| Model | Quant | Fits | Memory | Max context | Decode @32k | Prefill | Good for |
|---|---|---|---|---|---|---|---|
| Qwen 3.8 27B Alibaba Qwen | UD-Q5_K_M | fits | 22.0 GB | 37k65k with q8 cache | 10tok/sest. | 564 tok/s | – |
| Qwen 3.6 27B Alibaba Qwen | Q5_K_M | fits | 21.8 GB | 40k71k with q8 cache | 11tok/sest. | 564 tok/s | – |
| Gemma 4 31B | Q4_K_M | fits | 21.8 GB | 38k76k with q8 cache | 10tok/sest. | 502 tok/s | – |
| Qwen 3.6 35B-A3B Alibaba Qwen | MXFP4 | fits | 22.4 GB | 32k50k with q8 cache | 49tok/sest. | 4025 tok/s | chat, coding, agents, long documents |
| Qwen 3.5 27B Alibaba Qwen | Q5_K_M | fits | 21.9 GB | 39k69k with q8 cache | 10tok/sest. | 564 tok/s | – |
| Gemma 4 26B-A4B | UD-Q5_K_M | fits | 22.0 GB | 47k79k with q8 cache | 30tok/sest. | 3580 tok/s | chat, coding, agents, long documents |
| GLM 4.7 Flash Z.ai | Q4_K_M | fits | 20.2 GB | 70k121k with q8 cache | 26tok/sest. | 2101 tok/s | chat, coding |
| Devstral Small 2 24B Mistral AI | Q5_K_M | fits | 22.1 GB | 33k61k with q8 cache | 10tok/sest. | 595 tok/s | – |
| gpt-oss 20B OpenAI | MXFP4 | fits | 13.5 GB | 128k | 39tok/sest. | 3004 tok/s | chat, coding, agents, long documents |
| Qwen3 Coder 30B-A3B Alibaba Qwen | Q4_K_M | fits | 21.8 GB | 38k68k with q8 cache | 25tok/sest. | 1851 tok/s | chat, coding |
| Gemma 4 12B | Q8_0 | fits | 14.1 GB | 256k | 17tok/sest. | 1307 tok/s | chat, long documents |
| Qwen 3.5 9B Alibaba Qwen | Q8_0 | fits | 11.4 GB | 256k | 21tok/sest. | 1629 tok/s | chat, coding, long documents |
| Mistral Small 3.2 24B Mistral AI | Q5_K_M | fits | 22.1 GB | 33k61k with q8 cache | 10tok/sest. | 595 tok/s | – |
| Gemma 3 27B | Q5_K_M | fits | 22.3 GB | 32k61k with q8 cache | 10tok/sest. | 598 tok/s | – |
| Ministral 3 14B Mistral AI | Q8_0 | fits | 19.9 GB | 47k86k with q8 cache | 12tok/sest. | 905 tok/s | – |
| Qwen3 14B Alibaba Qwen | Q8_0 | fits | 21.1 GB | 39k40k with q8 cache | 11tok/sest. | 814 tok/s | – |
| DeepSeek R1 0528 Qwen3 8B DeepSeek | Q8_0 | fits | 14.1 GB | 88k128k with q8 cache | 17tok/sest. | 1342 tok/s | chat |
| Qwen3 8B Alibaba Qwen | Q8_0 | fits | 14.1 GB | 40k | 17tok/sest. | 1342 tok/s | chat |
| Phi-4 14B Microsoft | Q8_0 | fits | 22.2 GB | 16k | 10tok/sest. | 818 tok/s | – |
| Qwen 3.5 4B Alibaba Qwen | Q8_0 | fits | 6.7 GB | 256k | 41tok/sest. | 3048 tok/s | chat, coding, long documents |
| Ministral 3 8B Mistral AI | Q8_0 | fits | 14.1 GB | 91k166k with q8 cache | 17tok/sest. | 1296 tok/s | chat |
| Granite 4.2 8B IBM | Q8_0 | fits | 15.2 GB | 76k128k with q8 cache | 16tok/sest. | 1234 tok/s | chat |
| Gemma 4 E4B | Q8_0 | fits | 10.0 GB | 128k | 25tok/sest. | 2046 tok/s | chat, coding, long documents |
| LFM2.5 8B-A1B Liquid AI | Q8_0 | fits | 11.4 GB | 125k | 47tok/sest. | 6694 tok/s | chat, coding, long documents |
| Llama 3.1 8B Meta | Q6_K | fits | 11.6 GB | 114k128k with q8 cache | 21tok/sest. | 1418 tok/s | chat, coding |
| Qwen2.5 0.5B Alibaba Qwen | Q8_0 | fits | 2.4 GB | 32k | 244tok/sest. | 14579 tok/s | chat, coding, long documents |
| Qwen3.5 0.8B Alibaba Qwen | Q8_0 | fits | 2.6 GB | 256k | 185tok/sest. | 13694 tok/s | chat, coding, agents, long documents |
| Qwen3.5 2B Alibaba Qwen | Q5_K_M | fits | 3.2 GB | 256k | 123tok/sest. | 6529 tok/s | chat, coding, agents, long documents |
| DeepSeek R1 Distill Qwen 1.5B DeepSeek | Q8_0 | fits | 4.1 GB | 128k | 81tok/sest. | 5484 tok/s | chat, coding, long documents |
| MiniCPM5 2B openbmb | Q8_0 | fits | 5.3 GB | 128k | 56tok/sest. | 3275 tok/s | chat, coding, agents, long documents |
| LFM2.5 2.6B Liquid AI | Q8_0 | fits | 6.0 GB | 128k | 47tok/sest. | 3710 tok/s | chat, coding, agents, long documents |
| LFM2.5 VL 3B Liquid AI | Q8_0 | fits | 6.0 GB | 32k | 47tok/sest. | 3402 tok/s | chat, coding, agents, long documents |
| Mellum2.1 12B A2.5B Thinking JetBrains | Q8_0 | fits | 14.0 GB | 128k | 40tok/sest. | 4833 tok/s | chat, coding, agents, long documents |
| granite vision 4.1 4b IBM | Q8_0 | fits | 7.4 GB | 128k | 36tok/sest. | 2377 tok/s | chat, coding, long documents |
| granite 4.2 3b IBM | Q8_0 | fits | 7.6 GB | 128k | 35tok/sest. | 2491 tok/s | chat, coding, agents, long documents |
| LFM2.5 24B A2B Liquid AI | Q5_K_M | fits | 21.2 GB | 46k81k with q8 cache | 30tok/sest. | 3731 tok/s | chat, coding |
| Kimi VL A3B Moonshot AI | mlx-8bit | fits | 19.0 GB | 126k128k with q8 cache | 29tok/sest. | 3632 tok/s | chat, coding, long documents |
| DeepSeek R1 Distill Qwen 7B DeepSeek | Q8_0 | fits | 10.8 GB | 128k | 23tok/sest. | 1603 tok/s | chat, coding |
| GLM 4.6V Flash Z.ai | Q8_0 | fits | 12.1 GB | 128k | 20tok/sest. | 1116 tok/s | chat, coding |
| DeepSeek R1 Distill Llama 8B DeepSeek | Q8_0 | fits | 13.4 GB | 100k128k with q8 cache | 18tok/sest. | 1418 tok/s | chat |
| GLM 4.1V 9B Thinking Z.ai | mlx-8bit | fits | 13.7 GB | 64k | 17tok/sest. | 1116 tok/s | chat |
| Ministral 3 8B Reasoning 2512 Mistral AI | Q8_0 | fits | 14.1 GB | 91k166k with q8 cache | 17tok/sest. | 1296 tok/s | chat |
| DeepSeek R1 Distill Qwen 14B DeepSeek | Q8_0 | fits | 22.1 GB | 33k61k with q8 cache | 10tok/sest. | 766 tok/s | – |
| Devstral Small 2507 Mistral AI | Q5_K_M | fits | 22.1 GB | 33k61k with q8 cache | 10tok/sest. | 604 tok/s | – |
| Magistral Small 2506 Mistral AI | Q5_K_M | fits | 22.1 GB | 33k40k with q8 cache | 10tok/sest. | 604 tok/s | – |
| Magistral Small 2507 Mistral AI | Q5_K_M | fits | 22.1 GB | 33k40k with q8 cache | 10tok/sest. | 604 tok/s | – |
| Magistral Small 2509 Mistral AI | Q5_K_M | fits | 22.1 GB | 33k61k with q8 cache | 10tok/sest. | 595 tok/s | – |
| Qwen 3.5 122B-A10B Alibaba Qwen | UD-Q2_K_XL | offload | 41.3 GB | – | 12tok/sest. | 945 tok/s | – |
| Qwen3 Coder Next Alibaba Qwen | Q4_K_M | offload | 47.5 GB | – | 14tok/sest. | 2143 tok/s | chat |
| Mistral Small 4 119B Mistral AI | UD-Q2_K_XL | offload | 39.6 GB | – | 16tok/sest. | 1071 tok/s | chat |
| Qwen 3.5 35B-A3B Alibaba Qwen | Q4_K_M | offload | 22.7 GB | 21k33k with q8 cache | 47tok/sest. | 3993 tok/s | chat, coding, agents, long documents |
| Qwen3 Next 80B-A3B Alibaba Qwen | Q4_K_M | offload | 47.5 GB | – | 17tok/sest. | 2550 tok/s | chat |
| Nemotron 3.5 Lightning 30B-A3B NVIDIA | UD-Q6_K_XL | offload | 34.3 GB | – | 16tok/sest. | 3592 tok/s | chat |
| Kimi Linear 48B-A3B Moonshot AI | Q4_K_M | offload | 29.8 GB | – | 32tok/sest. | 2926 tok/s | chat, coding, long documents |
| Qwen3 32B Alibaba Qwen | Q4_K_M | offload | 27.9 GB | 10k19k with q8 cache | 5tok/sest. | 306 tok/s | – |
| DeepSeek R1 Distill Qwen 32B DeepSeek | Q4_K_M | offload | 28.0 GB | 10k18k with q8 cache | 5tok/sest. | 351 tok/s | – |
| granite 4.2 30b IBM | Q4_K_M | offload | 26.3 GB | 16k31k with q8 cache | 6tok/sest. | 418 tok/s | – |
| DeepSeek V4 Flash DeepSeek | UD-Q2_K_XL | won't run | 94.4 GB | – | – | – | – |
| GLM 5.3 Flash Z.ai | UD-Q2_K_XL | won't run | 103.2 GB | – | – | – | – |
| Qwen 3.8 Flash Next Alibaba Qwen | UD-Q2_K_XL | won't run | 75.8 GB | – | – | – | – |
| MiniMax M2.7 MiniMax | UD-Q2_K_XL | won't run | 79.4 GB | – | – | – | – |
| Nemotron 3 Super 120B-A12B NVIDIA | UD-Q2_K_XL | won't run | 52.8 GB | – | – | – | – |
| gpt-oss 120B OpenAI | Q4_K_M | won't run | 61.1 GB | – | – | – | – |
| GLM 4.5 Air Z.ai | Q2_K | won't run | 49.4 GB | – | – | – | – |
| Llama 3.3 70B Meta | Q4_K_M | won't run | 51.1 GB | – | – | – | – |
| DeepSeek R1 DeepSeek | Q4_K_M | won't run | 380.3 GB | – | – | – | – |
| DeepSeek R1 0528 DeepSeek | Q4_K_M | won't run | 380.8 GB | – | – | – | – |
| DeepSeek R1 Distill Llama 70B DeepSeek | Q4_K_M | won't run | 51.1 GB | – | – | – | – |
| DeepSeek V3 0324 DeepSeek | Q4_K_M | won't run | 380.3 GB | – | – | – | – |
| DeepSeek V3.1 DeepSeek | Q4_K_M | won't run | 381.2 GB | – | – | – | – |
| DeepSeek V3.2 DeepSeek | Q4_K_M | won't run | 381.2 GB | – | – | – | – |
| DeepSeek V4 Flash 0731 DeepSeek | MXFP4 | won't run | 149.8 GB | – | – | – | – |
| DeepSeek V4 Flash Vision Exp DeepSeek | MXFP4 | won't run | 158.6 GB | – | – | – | – |
| Devstral 2 123B Instruct 2512 Mistral AI | Q8_0 | won't run | 136.2 GB | – | – | – | – |
| GLM 4.5V Z.ai | Q4_K_M | won't run | 66.5 GB | – | – | – | – |
| GLM 4.6V Z.ai | Q4_K_M | won't run | 72.8 GB | – | – | – | – |
| GLM 4.7 Z.ai | Q4_K_M | won't run | 214.6 GB | – | – | – | – |
| GLM 5.1 Z.ai | MXFP4 | won't run | 424.2 GB | – | – | – | – |
| GLM 5.2 Z.ai | UD-Q4_K_M | won't run | 438.1 GB | – | – | – | – |
| GLM 5.3 Z.ai | Q8_0 | won't run | 750.6 GB | – | – | – | – |
| Kimi K2 Moonshot AI | Q4_K_M | won't run | 581.8 GB | – | – | – | – |
| Kimi K2 Instruct 0905 Moonshot AI | Q4_K_M | won't run | 582.2 GB | – | – | – | – |
| Kimi K2 Thinking Moonshot AI | Q4_K_M | won't run | 582.2 GB | – | – | – | – |
| Kimi K2.5 Moonshot AI | Q4_K_M | won't run | 582.2 GB | – | – | – | – |
| MiniMax M2 MiniMax | Q4_K_M | won't run | 138.1 GB | – | – | – | – |
| MiniMax M2.1 MiniMax | Q4_K_M | won't run | 138.1 GB | – | – | – | – |
| MiniMax M2.5 MiniMax | MXFP4 | won't run | 124.5 GB | – | – | – | – |
| MiniMax M3 MiniMax | UD-Q4_K_M | won't run | 252.0 GB | – | – | – | – |
| Mistral Medium 3.5 128B Mistral AI | Q4_K_M | won't run | 82.2 GB | – | – | – | – |
Specifications
From the manufacturer's published specifications; sources are linked where we have them.
- GPU
- Arc B580
- Memory
- 12 GB GDDR6
- Memory bus
- 192-bit
- Memory bandwidth
- 456 GB/s
- FP16 compute
- 116.5 TFLOPS
- Software
- SYCL / oneAPI
- Board power
- 190 W
- PCIe generation
- Gen 4
- PCIe lanes
- x8
Compare with
Questions
- How much memory does the Intel Arc B580 have for AI models?
- The Intel Arc B580 has 12 GB of GDDR6 memory with 456 GB/s of bandwidth, which sets how fast it generates text.
- What is the best model the Intel Arc B580 can run locally?
- Qwen 3.8 27B (UD-Q5_K_M) fits fully in its memory at 32k context and generates at about 10 tok/s, an estimate from its memory bandwidth. 47 of the 113 models we track fit at that context.
- Can the Intel Arc B580 run DeepSeek V4 Flash?
- No: even at UD-Q2_K_XL, DeepSeek V4 Flash needs about 94 GB at 32k context, more than the Intel Arc B580 and system RAM can hold.
- How much does the Intel Arc B580 cost in Canada?
- The best new price for the Intel Arc B580 in Canada is $550 at eBay; used, it starts at $450 (asking price).
Speeds on this page are for 32k context. See the best hardware for local AI at every budget.