FitMyLLM
INTEL· XE-HPG

Intel Arc A350

Running LLMs on the Arc A350 — the long read: which models fit at which quantisation, and the settings worth changing. · Or what else $99 buys

VRAM
4 GB
BUDGET
BANDWIDTH
124
GB/S
MODELS Q4
97/449
22%
7B Q4 SPEED
~6
SLOW
▸ MODEL COVERAGE @ Q422% OF ALL
▸ ESTIMATED SPEED· BY MODEL SIZE @ Q4

Average speeds at Q4 quantization. Actual performance varies by model architecture and context length.

3B
~13
TOK/S
7B
3.9GB NEEDED
14B
7.9GB NEEDED
32B
18.0GB NEEDED
70B
39.4GB NEEDED
▸ MEASURED RIG REPORTS

We rent the machine and time every model on it: decode, VRAM peak, concurrency, watts and cost per million tokens. Yours may already be one of them — and two are free to read in full.

SEE THE REPORTS
▸ SPECIFICATIONS
VRAM
4 GB
BANDWIDTH
124 GB/s
FP16 COMPUTE
6.1 TFLOPS
TDP
25W
MEMORY
GDDR6
ARCHITECTURE
Xe-HPG
COMPUTE UNITS
48
PCIE
Gen 4 x8
MSRP
$99
36
FAST MODELS · >30 TOK/S
Real-time chat speed
96
USABLE · >10 TOK/S
Comfortable for all tasks
97
TOTAL COMPATIBLE
Fit in VRAM at Q4
▸ DON’T WANT TO BUY?

Test Arc A350 (or anything bigger) without committing. Pay by the second, cancel anytime.

Spin up in ~60s. Pay by the second. Cancel anytime.

Some links are affiliate links — we may earn a small commission at no extra cost to you. This helps keep FitMyLLM free and independent.

▸ COMPATIBLE MODELS· 97
S
Falcon-H1R Tiny 90M0.09B
FALCON·256K CTX· CHAT· REASONING
441
TOK/S · 14% VRAM
S
SmolLM2 135M0.135B
SMOLLM·2K CTX· CHAT
294
TOK/S · 14% VRAM
S
nomic-embed-text-v1.5 100M0.14B
EMBEDDING·8K CTX· CHAT
283
TOK/S · 14% VRAM
S
GPT-2 124M0.14B
GPT2·1K CTX· CHAT
283
TOK/S · 14% VRAM
S
SmolVLM 256M0.256B
SMOLLM·8K CTX· CHAT· VISION
155
TOK/S · 16% VRAM
S
Gemma 3 270M0.27B
GEMMA·32K CTX· CHAT
147
TOK/S · 16% VRAM
S
Snowflake Arctic Embed M v2.00.305B
EMBEDDING·8K CTX· EMBEDDING· MULTILINGUAL
130
TOK/S · 17% VRAM
S
bge-large-en-v1.5 335M0.335B
EMBEDDING·1K CTX· CHAT
118
TOK/S · 17% VRAM
S
mxbai-embed-large-v10.335B
EMBEDDING·1K CTX· EMBEDDING
118
TOK/S · 17% VRAM
S
Snowflake Arctic Embed L0.335B
EMBEDDING·1K CTX· EMBEDDING
118
TOK/S · 17% VRAM
S
LFM2 350M0.35B
LFM·125K CTX· CHAT· TOOL_USE
113
TOK/S · 18% VRAM
S
SmolLM2 360M0.36B
SMOLLM·8K CTX· CHAT
110
TOK/S · 18% VRAM
S
GPT-2 Medium 345M0.38B
GPT2·1K CTX· CHAT
104
TOK/S · 18% VRAM
S
Qwen 2.5 0.5B0.5B
QWEN·32K CTX· CHAT
79
TOK/S · 20% VRAM
S
SmolVLM 500M0.5B
SMOLLM·8K CTX· CHAT· VISION
79
TOK/S · 20% VRAM
S
Falcon-H1 0.5B0.52B
FALCON·128K CTX· CHAT
76
TOK/S · 20% VRAM
S
BGE-M30.568B
EMBEDDING·8K CTX· EMBEDDING
70
TOK/S · 21% VRAM
S
Snowflake Arctic Embed L v2.00.568B
EMBEDDING·8K CTX· EMBEDDING· MULTILINGUAL
70
TOK/S · 21% VRAM
S
Qwen3-Embedding 0.6B0.6B
EMBEDDING·32K CTX· EMBEDDING· MULTILINGUAL
66
TOK/S · 21% VRAM
S
Falcon-H1R Tiny 0.6B0.6B
FALCON·32K CTX· CHAT· REASONING
66
TOK/S · 21% VRAM
S
Falcon Perception 0.6B0.6B
FALCON·4K CTX· VISION
66
TOK/S · 21% VRAM
S
Qwen 1.5 0.5B0.62B
QWEN·32K CTX· CHAT
64
TOK/S · 22% VRAM
A
LFM2 700M0.74B
LFM·125K CTX· CHAT· TOOL_USE
54
TOK/S · 24% VRAM
A
Qwen3 0.6B0.75B
QWEN·32K CTX· CHAT· REASONING
53
TOK/S · 24% VRAM
A
GPT-2 Large 774M0.81B
GPT2·1K CTX· CHAT
49
TOK/S · 25% VRAM
A
Qwen 3.5 0.8B0.87B
QWEN·256K CTX· CHAT· CODING· MULTILINGUAL
46
TOK/S · 26% VRAM
A
InternVL3 1B0.94B
OTHER·32K CTX· CHAT· VISION
42
TOK/S · 27% VRAM
B
MiniCPM5 1B1.08B
MINICPM·128K CTX· CHAT· REASONING· MULTILINGUAL
37
TOK/S · 29% VRAM
B
TinyLlama 1.1B1.1B
LLAMA·2K CTX· CHAT
36
TOK/S · 29% VRAM
B
LFM2.5-1.2B-Thinking1.2B
LFM·122K CTX· CHAT· REASONING· TOOL_USE
33
TOK/S · 31% VRAM
B
Llama-3.2-1B1.2B
LLAMA·4K CTX· CHAT
33
TOK/S · 31% VRAM
B
LFM2 1.2B1.2B
LFM·125K CTX· CHAT· TOOL_USE· MULTILINGUAL
33
TOK/S · 31% VRAM
B
Zamba2 1.2B1.2B
OTHER·4K CTX· CHAT
33
TOK/S · 31% VRAM
B
EXAONE-4.0-1.2B1.3B
EXAONE·64K CTX· CHAT
31
TOK/S · 32% VRAM
B
OPT 1.3B1.3B
OPT·2K CTX· CHAT
31
TOK/S · 32% VRAM
B
MiniCPM-V 4.61.3B
OTHER·256K CTX· CHAT· VISION
31
TOK/S · 32% VRAM
B
DeepSeek Coder 1.3B1.35B
DEEPSEEK·16K CTX· CODING
29
TOK/S · 33% VRAM
B
Phi-1 1.3B1.42B
PHI·2K CTX· CODING
28
TOK/S · 34% VRAM
B
Phi-1.5 1.3B1.42B
PHI·2K CTX· CHAT· CODING
28
TOK/S · 34% VRAM
B
Falcon-H1 1.5B1.55B
FALCON·128K CTX· CHAT· CODING
26
TOK/S · 36% VRAM
▸ NEXT STEP

Get personalized recommendations.

See ranked models with benchmark scores, run commands, and precise speed estimates for your Arc A350.

WHAT THIS CARD IS WORTH

Arc A350 holds 97 of the models in our catalogue and is, in practice, a Q4_K_M card — the largest it takes is InternLM2 5B at Q4_K_M.

TOKENS/SEC PER $100
8B at Q4_K_M, so cards compare like for like
VRAM PER $100
4.04 GB
what memory costs on this card
THE BIGGEST IT TAKES
InternLM2 5B4.5B · Q4_K_M3.6 GB8 tok/sEST
Ministral 3 3B Reasoning4.25B · Q4_K_M3.3 GB9 tok/sEST
Qwen3 4B4B · Q4_K_M3.3 GB9 tok/sEST
Qwen3-4B Instruct 25074B · Q4_K_M3.3 GB9 tok/sEST
Qwen3-Embedding 4B4B · Q4_K_M3.3 GB9 tok/sEST
Nemotron 3 Nano 4B3.97B · Q4_K_M3.4 GB9 tok/sEST
Ministral 3 3B3.85B · Q5_K_M3.5 GB8 tok/sEST
phi-3-mini-4k 3.8B3.8B · Q4_K_M3.4 GB10 tok/sEST
AGAINST CARDS OF SIMILAR MEMORY
RTX 3070 Ti8 GB$49912.6 tok/s per $100
RTX 3060 Ti GDDR6X8 GB$39915.8 tok/s per $100
RTX 3070 Ti 8 GB GA1028 GB$59910.5 tok/s per $100
Arc A7508 GB$19910.1 tok/s per $100
Arc A3504 GB$99.00

Neighbours in memory rather than in price: memory decides whether a card can do the job at all, so two cards of the same size at different prices is the comparison you are making. Ordered by memory, then bandwidth — not by the value column, which is worked out from bandwidth and price alone and therefore rewards a cheap card whatever its software stack does to that bandwidth in practice. Read it as one input, not as a ranking.

▸ DEVICE UNDER TEST

Intel Arc A350 — 4 GB VRAM.

ARC A350 SPEC
BRAND
Intel
VRAM
4 GB GDDR6
BANDWIDTH
124 GB/s
FP16 COMPUTE
6.1 TFLOPS
FP32 COMPUTE
3.1 TFLOPS
TDP
25 W
ARCHITECTURE
Xe-HPG
MSRP
$99
▸ AI CAPABILITY
97/ 449 models @ Q4

With 4 GB VRAM and 124 GB/s bandwidth, this GPU handles models up to 4.25B parameters.

Speed ≈ bandwidth / model_size × efficiency. A 7B model at Q4 runs at ~14 tok/s.

§ 01TOP MODELS FOR ARC A350
97 FIT · SHOWING 20
MODELSIZEVRAM Q4TOK/SAVG
Ministral 3 3B Reasoning4.25B3.1 GB9
Qwen 1.5 4B4B2.9 GB1012.6
Qwen3 4B4B2.9 GB1040.7
Qwen3-4B Instruct 25074B2.9 GB1037.2
Qwen3-Embedding 4B4B2.9 GB10
Nemotron 3 Nano 4B3.97B2.9 GB1032.0
Ministral 3 3B3.85B2.8 GB1021.4
Phi-3.5 Mini 3.8B3.82B2.8 GB1046.6
phi-3-mini-4k 3.8B3.8B2.8 GB1030.5
Phi-4-mini 3.8B3.8B2.8 GB1049.0
Cogito 3B3.61B2.7 GB1122.1
Falcon3-3B3.23B2.5 GB1225.7
granite-4.0-h-micro 3.2B3.2B2.4 GB1218.4
Llama-3.2-3B3.2B2.4 GB1217.9
Falcon-H1 3B3.15B2.4 GB1349.5
Qwen 2.5 3B3.1B2.4 GB1337.2
SmolLM3-3B3.1B2.4 GB1330.5
Ministral 3B3B2.3 GB1329.6
StarCoder2 3B3B2.3 GB139.5
Granite 4.1 3B3B2.3 GB1316.6