FitMyLLM
AMD· RDNA 3.0

AMD Radeon PRO V710

Running LLMs on the Radeon PRO V710 — the long read: which models fit at which quantisation, and the settings worth changing. · Or what a budget buys

VRAM
28 GB
High-end
Bandwidth
504
GB/s
Models Q4
357/498
72%
7B Q4 speed
~64
Blazing
Model coverage at Q472% of all
Estimated speed· By model size at Q4

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

3B
~149
tok/s
7B
~64
tok/s
14B
~32
tok/s
32B
~14
tok/s
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
28 GB
Bandwidth
504 GB/s
FP16 Compute
55.3 TFLOPS
TDP
158W
Memory
GDDR6
Architecture
RDNA 3.0
Stream Processors
3,456
Compute Units
54
PCIe
Gen 4 x16
295
Fast models · >30 tok/s
Real-time chat speed
357
Usable · >10 tok/s
Comfortable for all tasks
357
Total compatible
Fit in VRAM at Q4
Don’t want to buy?

Test Radeon PRO V710 (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· 357
S
Falcon-H1R Tiny 90M0.09B
Falcon·256K ctx· chat· reasoning
4978
tok/s · 2% VRAM
›
S
SmolLM2 135M0.135B
Smollm·2K ctx· chat
3319
tok/s · 2% VRAM
›
S
nomic-embed-text-v1.5 100M0.14B
Embedding·8K ctx· chat
3200
tok/s · 2% VRAM
›
S
GPT-2 124M0.14B
Gpt2·1K ctx· chat
3200
tok/s · 2% VRAM
›
S
SmolVLM 256M0.256B
Smollm·8K ctx· chat· vision
1750
tok/s · 2% VRAM
›
S
Gemma 3 270M0.27B
Gemma·32K ctx· chat
1659
tok/s · 2% VRAM
›
S
Snowflake Arctic Embed M v2.00.305B
Embedding·8K ctx· embedding· multilingual
1469
tok/s · 2% VRAM
›
S
bge-large-en-v1.5 335M0.335B
Embedding·1K ctx· chat
1337
tok/s · 2% VRAM
›
S
mxbai-embed-large-v10.335B
Embedding·1K ctx· embedding
1337
tok/s · 2% VRAM
›
S
Snowflake Arctic Embed L0.335B
Embedding·1K ctx· embedding
1337
tok/s · 2% VRAM
›
S
LFM2 350M0.35B
Lfm·125K ctx· chat· tool_use
1280
tok/s · 3% VRAM
›
S
SmolLM2 360M0.36B
Smollm·8K ctx· chat
1244
tok/s · 3% VRAM
›
S
GPT-2 Medium 345M0.38B
Gpt2·1K ctx· chat
1179
tok/s · 3% VRAM
›
S
Qwen 2.5 0.5B0.5B
Qwen·32K ctx· chat
896
tok/s · 3% VRAM
›
S
SmolVLM 500M0.5B
Smollm·8K ctx· chat· vision
896
tok/s · 3% VRAM
›
S
Falcon-H1 0.5B0.52B
Falcon·128K ctx· chat
862
tok/s · 3% VRAM
›
S
BGE-M30.568B
Embedding·8K ctx· embedding
789
tok/s · 3% VRAM
›
S
Snowflake Arctic Embed L v2.00.568B
Embedding·8K ctx· embedding· multilingual
789
tok/s · 3% VRAM
›
S
Qwen3-Embedding 0.6B0.6B
Embedding·32K ctx· embedding· multilingual
747
tok/s · 3% VRAM
›
S
Falcon-H1R Tiny 0.6B0.6B
Falcon·32K ctx· chat· reasoning
747
tok/s · 3% VRAM
›
S
Falcon Perception 0.6B0.6B
Falcon·4K ctx· vision
747
tok/s · 3% VRAM
›
S
Qwen 1.5 0.5B0.62B
Qwen·32K ctx· chat
723
tok/s · 3% VRAM
›
S
LFM2 700M0.74B
Lfm·125K ctx· chat· tool_use
605
tok/s · 3% VRAM
›
S
Qwen3 0.6B0.75B
Qwen·32K ctx· chat· reasoning
597
tok/s · 3% VRAM
›
S
GPT-2 Large 774M0.81B
Gpt2·1K ctx· chat
553
tok/s · 4% VRAM
›
S
Qwen 3.5 0.8B0.87B
Qwen·256K ctx· chat· coding· multilingual
515
tok/s · 4% VRAM
›
S
InternVL3 1B0.94B
Other·32K ctx· chat· vision
477
tok/s · 4% VRAM
›
S
MiniCPM5 1B1.08B
Minicpm·128K ctx· chat· reasoning· multilingual
415
tok/s · 4% VRAM
›
S
K2 Horizon 0.9B1.08B
Other·128K ctx· chat· coding· reasoning
415
tok/s · 4% VRAM
›
S
TinyLlama 1.1B1.1B
Llama·2K ctx· chat
407
tok/s · 4% VRAM
›
S
LFM2.5-1.2B-Thinking1.2B
Lfm·122K ctx· chat· reasoning· tool_use
373
tok/s · 4% VRAM
›
S
Llama-3.2-1B1.2B
Llama·4K ctx· chat
373
tok/s · 4% VRAM
›
S
LFM2 1.2B1.2B
Lfm·125K ctx· chat· tool_use· multilingual
373
tok/s · 4% VRAM
›
S
Zamba2 1.2B1.2B
Other·4K ctx· chat
373
tok/s · 4% VRAM
›
S
Ling 3.0 Tiny7.89BMoE
Other·128K ctx· chat· coding· reasoning
345
tok/s · 19% VRAM
›
S
EXAONE-4.0-1.2B1.3B
Exaone·64K ctx· chat
345
tok/s · 5% VRAM
›
S
OPT 1.3B1.3B
Opt·2K ctx· chat
345
tok/s · 5% VRAM
›
S
MiniCPM-V 4.61.3B
Other·256K ctx· chat· vision
345
tok/s · 5% VRAM
›
S
DeepSeek Coder 1.3B1.35B
Deepseek·16K ctx· coding
332
tok/s · 5% VRAM
›
S
Phi-1 1.3B1.42B
Phi·2K ctx· coding
315
tok/s · 5% VRAM
›
Next step

Get personalized recommendations.

See ranked models with benchmark scores, run commands, and precise speed estimates for your Radeon PRO V710.

What this card is worth

Radeon PRO V710 holds 357 of the models in our catalogue and is, in practice, a IQ4_XS card — the largest it takes is Flint at IQ4_XS.

Tokens/sec per $100
—
8B at Q4_K_M, so cards compare like for like
VRAM per $100
—
what memory costs on this card
The biggest it takes
Flint39.79B · IQ4_XS24.8 GB134 tok/sest
InternVL3 38B38B · IQ4_XS24.7 GB16 tok/sest
K2 Horizon MoVA 36B-A4B37.44B · Q4_K_S25.1 GB92 tok/sest
Seed-OSS 36B Instruct36B · Q4_K_S24.5 GB16 tok/sest
MiniCPM-V 4.7 35B-A3B35.21B · Q4_K_M24.1 GB110 tok/sest
c4ai-command-r-v01 35B35B · Q4_K_M24.5 GB16 tok/sest
Qwen 3.5 35B A3B35B · Q4_K_M24.0 GB110 tok/sest
Qwen 3.6 35B A3B35B · Q4_K_M24.0 GB110 tok/sest
Against cards of similar memory
M4 Max (48GB)36 GB$2,4992.2 tok/s per $100
M3 Max (48GB)36 GB$2,8991.6 tok/s per $100
M4 Pro (48GB)36 GB$1,7991.9 tok/s per $100
M3 Max (36GB)27 GB$2,4991.5 tok/s per $100
Radeon PRO V71028 GB——

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

AMD Radeon PRO V710 — 28 GB VRAM.

Radeon PRO V710 spec
Brand
AMD
VRAM
28 GB GDDR6
Bandwidth
504 GB/s
FP16 Compute
55.3 TFLOPS
FP32 Compute
27.6 TFLOPS
Stream Processors
3,456
TDP
158 W
Architecture
RDNA 3.0
AI capability
357/ 498 models @ Q4

With 28 GB VRAM and 504 GB/s bandwidth, this GPU handles models up to 36B parameters.

Speed ≈ bandwidth / model_size × efficiency. A 7B model at Q4 runs at ~58 tok/s.

Rent it — live prices
Checking what the clouds are charging…
Top models for Radeon PRO V710
357 fit · showing 20
ModelSizeVRAM Q4tok/sAvg
Seed-OSS 36B Instruct36B22.5 GB1254.4
MiniCPM-V 4.7 35B-A3B35.21B22.0 GB14976.4
c4ai-command-r-v01 35B35B21.9 GB1327.5
Qwen 3.5 35B A3B35B21.9 GB14956.1
Qwen 3.6 35B A3B35B21.9 GB14953.3
K2 Horizon 32B34.78B21.7 GB1354.0
Nous Capybara 34B34.4B21.5 GB1342.0
Yi-1.5 34B34.4B21.5 GB1345.3
Falcon-H1 34B34B21.3 GB1358.6
CodeLlama 34B34B21.3 GB1325.4
Nous Hermes 2 34B34B21.3 GB1347.0
Phind CodeLlama 34B34B21.3 GB1368.1
LLaVA-1.6 Yi 34B34B21.3 GB1347.4
WizardCoder Python 34B34B21.3 GB1373.2
Yi 34B34B21.3 GB1333.4
Qwen3-VL 32B Instruct33.36B20.9 GB1344.6
Nanbeige4.3 33B33.12B20.7 GB1492.4
DeepSeek Coder 33B33B20.7 GB1426.0
Vicuna 33B33B20.7 GB1417.2
LLaMA 1 30B33B20.7 GB1417.8