Intel · Alchemist · 8 GB VRAM tier

Best AI Models for Intel Arc A580 (8 GB)

Runs 7–9B models well; 14B doesn't fit in 8 GB.

8 GB GDDR6512 GB/s99 TFLOPS FP16185 WAlchemist 2023

8 GB fits ~7-8B Q4 models. Runs through llama.cpp's Vulkan or SYCL backends; Intel's IPEX-LLM is no longer maintained.

What LLMs can the Arc A580 run?

#	Model	Speed	Quant · VRAM
1	Qwen3.5 9BAlibabaHow to run → Q4_K_M · 6.6 GB · 40.3 t/s	40.3 tok/s	Q4_K_M · 6.6 GB
2	Qwen3.5 4BAlibaba Q4_K_M · 3.4 GB · 78.5 t/s	78.5 tok/s	Q4_K_M · 3.4 GB
3	Gemma 2 9B ITGoogle Q4_K_M · 7.3 GB · 42.0 t/s	42.0 tok/s	Q4_K_M · 7.3 GB
4	Gemma 3 4B ITGoogle Q4_K_M · 5.0 GB · 90.1 t/s	90.1 tok/s	Q4_K_M · 5.0 GB
5	Ling 3.0 TinyInclusion AI Q4_K_M · 5.5 GB · 118 t/s	118 tok/s	Q4_K_M · 5.5 GB

#	Model	Speed	Reasoning Score	Quant · VRAM
1	Qwen3.5 9BAlibabaHow to run → Q4_K_M · 6.6 GB · 40.3 t/s	40.3 tok/s		Q4_K_M · 6.6 GB
2	Gemma 3 4B ITGoogle Q4_K_M · 5.0 GB · 90.1 t/s	90.1 tok/s		Q4_K_M · 5.0 GB
3	Qwen2.5 7B InstructAlibaba Q4_K_M · 5.8 GB · 51.3 t/s	51.3 tok/s		Q4_K_M · 5.8 GB
4	Gemma 2 9B ITGoogle Q4_K_M · 7.3 GB · 42.0 t/s	42.0 tok/s	overall	Q4_K_M · 7.3 GB
5	Ling 3.0 TinyInclusion AI Q4_K_M · 5.5 GB · 118 t/s	118 tok/s		Q4_K_M · 5.5 GB

#	Model	Speed	Quant · VRAM
1	Qwen3.5 4BAlibabaHow to run → Q4_K_M · 3.4 GB · 78.5 t/s	78.5 tok/s	Q4_K_M · 3.4 GB
2	Qwen3.5 9BAlibaba Q4_K_M · 6.6 GB · 40.3 t/s	40.3 tok/s	Q4_K_M · 6.6 GB
3	Gemma 4 E4B ITGoogle Q4_K_M · 5.6 GB · 48.1 t/s	48.1 tok/s	Q4_K_M · 5.6 GB
4	Ling 3.0 TinyInclusion AI Q4_K_M · 5.5 GB · 118 t/s	118 tok/s	Q4_K_M · 5.5 GB
5	Spark X2.6 4BXHToken Q4_K_M · 4.3 GB · 86.5 t/s	86.5 tok/s	Q4_K_M · 4.3 GB

#	Model	Speed ↓	Quant · VRAM
1	Qwen3.5 4BAlibabaHow to run → Q4_K_M · 3.4 GB · 78.5 t/s	78.5 tok/s	Q4_K_M · 3.4 GB
2	Ling 3.0 TinyInclusion AI Q4_K_M · 5.5 GB · 118 t/s	118 tok/s	Q4_K_M · 5.5 GB
3	Gemma 4 E4B ITGoogle Q4_K_M · 5.6 GB · 48.1 t/s	48.1 tok/s	Q4_K_M · 5.6 GB
4	MiMo V2.6 Distill Qwen 9BXiaomi Q4_K_M · 6.5 GB · 41.2 t/s	41.2 tok/s	Q4_K_M · 6.5 GB
5	Qwen3.5 9BAlibaba Q4_K_M · 6.6 GB · 40.3 t/s	40.3 tok/s	Q4_K_M · 6.6 GB

#	Model	Speed	Quant · VRAM
1	Qwen3.5 9BAlibabaHow to run → Q4_K_M · 6.6 GB · 40.3 t/s	40.3 tok/s	Q4_K_M · 6.6 GB
2	Qwen3.5 4BAlibaba Q4_K_M · 3.4 GB · 78.5 t/s	78.5 tok/s	Q4_K_M · 3.4 GB
3	Gemma 3 4B ITGoogle Q4_K_M · 5.0 GB · 90.1 t/s	90.1 tok/s	Q4_K_M · 5.0 GB
4	MiMo V2.6 Distill Qwen 9BXiaomi Q4_K_M · 6.5 GB · 41.2 t/s	41.2 tok/s	Q4_K_M · 6.5 GB
5	MiniCPM V 4.7 1BOpenBMB Q4_K_M · 1.6 GB · 200 t/s	200 tok/s	Q4_K_M · 1.6 GB

See all 218 models on the Arc A580 →Best quant that fits · speeds estimated on this card

About the Arc A580

The Arc A580 is an entry card in the 8 GB class: it runs 7–9B models well. It is more efficient per watt than most 8 GB cards. Alchemist generation, released 2023.

Speed in 8 GB#11512 GB/s

Per watt#90.22 tok/s per W on Qwen3.5 9B (9B params)

GenerationAlchemistreleased 2023

Arc A580 strengths and weaknesses

Strengths

Runs 7–9B models: Qwen3.5 9B at 40.3 tok/s in 6.6 GB

Weaknesses

14B models don't fit: Qwen3 14B needs 11.9 GB at Q4 vs 8 GB here
Little room for context: Qwen3.5 9B uses 6.6 of 8 GB, 1.4 GB left
oneAPI, not CUDA: No FP8 or NVFP4 weights and no vLLM AWQ or GPTQ; use GGUF builds

Speed per dollar and per watt

NewUsed

On Qwen3.5 9B (9B params) at Q4_K_M. New prices are compared with new, used with used. The last two columns are the speed you get per $100 of price and per 100 W of board power; higher is better. The Arc A580 has no current price, so it is shown at its MSRP.

Card	New price	tok/s	tok/s per $100	tok/s per 100 W
Arc A5808 GB	about $179	40.3	22.5	21.8
RTX 50608 GBBest new value	about $460	45.8	10.0	31.6
RTX 507012 GB	about $840	68.7	8.2	27.5
RTX 40608 GB	about $520	27.8	5.3	24.2

Card	Used price	tok/s	tok/s per $100	tok/s per 100 W
RTX 3060 12GB12 GB	about $329	36.8	11.2	21.6
RTX 50608 GBBest used value	about $430	45.8	10.7	31.6
RTX 507012 GB	about $736	68.7	9.3	27.5
RTX 407012 GB	about $575	51.5	9.0	25.8
RTX 40608 GB	about $335	27.8	8.3	24.2
RTX 4070 SUPER12 GB	about $675	51.5	7.6	23.4
RTX 4060 Ti 8GB8 GB	about $400	29.4	7.4	18.4
RTX 4070 Ti12 GB	about $722	51.5	7.1	18.1

Cost per token: the Arc A580 against 8–12 GB cards

On Qwen3.5 9B (9B params, Q4_K_M, 4K context). Speed, cost per token and power per token depend on the model.

8 GB12 GBthis cardat MSRP

Include card price

Power$/kWhUseh/dayKeepyears

Each dot is a card; hover or tap it for its numbers, click to open it. Cards without a price for the chosen basis are left out. How we calculate cost per token

Similar GPUs

Cards people weigh against the Arc A580, each with the one reason it is worth a look. Speeds on Qwen3.5 9B.

14% faster, same 8 GBRTX 5060NVIDIA · 8 GB · Blackwell

▲ 14% speedsame 8 GB$460

4 GB more and 71% fasterRTX 5070NVIDIA · 12 GB · Blackwell

▲ 71% speed▲ 4 GB$840rent $0.14/hr

4 GB more, but slowerRTX 3060 12GBNVIDIA · 12 GB · Ampere

▼ 9% speed▲ 4 GB$329rent $0.06/hr

8 GB more and 82% fasterRTX 4080NVIDIA · 16 GB · Ada Lovelace

▲ 82% speed▲ 8 GB$1,199rent $0.25/hr

8 GB more and nearly 2× the speedRTX 4080 SUPERNVIDIA · 16 GB · Ada Lovelace

▲ 87% speed▲ 8 GB$999rent $0.23/hr

2 GB more and nearly 2× the speedRTX 3080NVIDIA · 10 GB · Ampere

▲ 93% speed▲ 2 GB$699rent $0.11/hr

Arc A580 tokens per second by model size

Output speed depends on memory bandwidth (512 GB/s). Grey rows need more than 8 GB.

9BQwen3.5 9B · Q4_K_M · 6.6 GB
40.3tok/s
14BQwen3 14B · not at 4-bit
needs 11.9 GB
27–31BQwen3.8 27B · not at 4-bit
needs 17.9 GB
35B MoEQwen3.6 35B A3B · not at 4-bit
needs 22.0 GB
70BLlama 3.3 70B Instruct · not at 4-bit
needs 51.3 GB
120B MoEGPT OSS 120B · not at 4-bit
needs 70.8 GB

Estimated speeds: Q4_K_M, 4K context, one user. How we estimate

All 842 models on the Arc A580

Featured only

Best quant that fits, ordered by llmrun Score. Fit grades how well it runs here, S best. How ranking works

LLM models compatible with Intel Arc A580 — ranked by Score, recency and Fit
Model	Quant	VRAM	Speed	Context	Fit
Nemotron Flash 3B2.7BChat BF16·42.5 t/s	BF16	7.6 GB95%	42.5 t/s	29K	B53, Decent
Moonlight 16B A3B Instruct16.0BChat IQ3_M·90.4 t/s	IQ3_M	7.7 GB97%	90.4 t/s	8K	B62, Decent
EuroMoE 2.6B A0.6B 25122.6BChat Q4_K_M·160.5 t/s	Q4_K_M	2.0 GB25%	160.5 t/s	4K	S97, Runs great
Gemma 2 2B IT2.6BChat Q4_K_M·148.0 t/s	Q4_K_M	2.5 GB32%	148.0 t/s	8K	S97, Runs great
Tiny Qwen2ForCausalLM 2.52MChatFunctions Q4_K_M·853.3 t/s	Q4_K_M	0.6 GB7%	853.3 t/s	33K	S100, Runs great
Nemotron Content Safety Reasoning 4B4.3BChatReasoning Q4_K_M·79.0 t/s	Q4_K_M	3.6 GB45%	79.0 t/s	131K	S87, Runs great
MiniCPM O 2 68.7BVisionFunctions Q4_K_M·45.6 t/s	Q4_K_M	6.4 GB81%	45.6 t/s	33K	A76, Runs well
Qwen3 14B PARO1.6BChat Q4_K_M·163.1 t/s	Q4_K_M	3.9 GB49%	163.1 t/s	41K	S97, Runs great
Gemma 2 Mitra E9.2BChat Q4_K_M·39.6 t/s	Q4_K_M	7.7 GB96%	39.6 t/s	8K	B50, Decent
Saul 7B Instruct V17.2BChat Q4_K_M·52.1 t/s	Q4_K_M	6.8 GB85%	52.1 t/s	33K	A74, Runs well
Granite 4.0 Tiny Preview6.7BChatFunctions Q4_K_M·86.0 t/s	Q4_K_M	5.6 GB71%	86.0 t/s	131K	S88, Runs great
Orca Mini 3B3.4BChatMath Q4_K_M·113.3 t/s	Q4_K_M	2.3 GB28%	113.3 t/s	2K	S94, Runs great
Falcon 11B11.1BChat IQ4_XS·39.0 t/s	IQ4_XS	7.8 GB97%	39.0 t/s	8K	B48, Decent
C4ai Command R7b 12 20248.0BChatFunctions Q4_K_M·48.3 t/s	Q4_K_M	7.9 GB98%	48.3 t/s	—	C44, Usable
YanoljaNEXT EEVE Instruct 2.8B2.8BChat Q4_K_M·96.2 t/s	Q4_K_M	2.7 GB33%	96.2 t/s	2K	S90, Runs great

Arc A580 specs, power and PSU

Specs

Architecture: Alchemist
VRAM: 8 GB GDDR6
Memory bandwidth: 512.0 GB/s
FP16 (dense): 98.50 TFLOPS
Stream processors: 3,072
Tensor cores: 384
TDP: 185 W
Released: 2023-10-10
MSRP: $179

Arc A580 FAQ

What is the largest model the Arc A580 can run?

Qwen3.5 9B (9.7B) is the largest at 4-bit or better: at Q4_K_M it uses 6.6 GB of 8 GB, and runs at about 40.3 tok/s.

Can the Arc A580 run 14B models?

No. Qwen3 14B needs 11.9 GB at Q4, more than the 8 GB on this card.

How many tokens per second does the Arc A580 give?

About 40.3 tok/s on Qwen3.5 9B at Q4_K_M, with the first token after 4.5 s on a 4K-token prompt.

What power supply does the Arc A580 need?

Plan a 400 W PSU (2× the 185 W TDP).

Is renting cheaper than buying an Arc A580?

An RTX 5080 (16 GB) rents from $0.23/hr. Renting costs as much as buying after ≈800 hours (≈0.5 years at 4 h a day).

Best coding LLMs 8 GB VRAM 12 GB VRAM GPUs with 8 GB or more RTX 5080 RTX 3080 Ti