FitMyLLM
▸ AMD· GCN 4.0

AMD Radeon Pro 580

Running LLMs on the Radeon Pro 580 — the long read: which models fit at which quantisation, and the settings worth changing. · Or what a budget buys

VRAM
8 GB
ENTRY-LEVEL
BANDWIDTH
217
GB/S
MODELS Q4
201/449
45%
7B Q4 SPEED
~28
GOOD
▸ MODEL COVERAGE @ Q445% OF ALL
▸ ESTIMATED SPEED· BY MODEL SIZE @ Q4

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

3B
~64
TOK/S
7B
~28
TOK/S
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
8 GB
BANDWIDTH
217 GB/s
FP16 COMPUTE
5.5 TFLOPS
TDP
185W
MEMORY
GDDR5
ARCHITECTURE
GCN 4.0
STREAM PROCESSORS
2,304
COMPUTE UNITS
36
PCIE
Gen 3 x16
115
FAST MODELS · >30 TOK/S
Real-time chat speed
201
USABLE · >10 TOK/S
Comfortable for all tasks
201
TOTAL COMPATIBLE
Fit in VRAM at Q4
▸ DON’T WANT TO BUY?

Test Radeon Pro 580 (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· 201
S
Falcon-H1R Tiny 90M0.09B
FALCON·256K CTX· CHAT· REASONING
2143
TOK/S · 7% VRAM
›
S
SmolLM2 135M0.135B
SMOLLM·2K CTX· CHAT
1429
TOK/S · 7% VRAM
›
S
nomic-embed-text-v1.5 100M0.14B
EMBEDDING·8K CTX· CHAT
1378
TOK/S · 7% VRAM
›
S
GPT-2 124M0.14B
GPT2·1K CTX· CHAT
1378
TOK/S · 7% VRAM
›
S
SmolVLM 256M0.256B
SMOLLM·8K CTX· CHAT· VISION
753
TOK/S · 8% VRAM
›
S
Gemma 3 270M0.27B
GEMMA·32K CTX· CHAT
714
TOK/S · 8% VRAM
›
S
Snowflake Arctic Embed M v2.00.305B
EMBEDDING·8K CTX· EMBEDDING· MULTILINGUAL
632
TOK/S · 8% VRAM
›
S
bge-large-en-v1.5 335M0.335B
EMBEDDING·1K CTX· CHAT
576
TOK/S · 9% VRAM
›
S
mxbai-embed-large-v10.335B
EMBEDDING·1K CTX· EMBEDDING
576
TOK/S · 9% VRAM
›
S
Snowflake Arctic Embed L0.335B
EMBEDDING·1K CTX· EMBEDDING
576
TOK/S · 9% VRAM
›
S
LFM2 350M0.35B
LFM·125K CTX· CHAT· TOOL_USE
551
TOK/S · 9% VRAM
›
S
SmolLM2 360M0.36B
SMOLLM·8K CTX· CHAT
536
TOK/S · 9% VRAM
›
S
GPT-2 Medium 345M0.38B
GPT2·1K CTX· CHAT
508
TOK/S · 9% VRAM
›
S
Qwen 2.5 0.5B0.5B
QWEN·32K CTX· CHAT
386
TOK/S · 10% VRAM
›
S
SmolVLM 500M0.5B
SMOLLM·8K CTX· CHAT· VISION
386
TOK/S · 10% VRAM
›
S
Falcon-H1 0.5B0.52B
FALCON·128K CTX· CHAT
371
TOK/S · 10% VRAM
›
S
BGE-M30.568B
EMBEDDING·8K CTX· EMBEDDING
340
TOK/S · 10% VRAM
›
S
Snowflake Arctic Embed L v2.00.568B
EMBEDDING·8K CTX· EMBEDDING· MULTILINGUAL
340
TOK/S · 10% VRAM
›
S
Qwen3-Embedding 0.6B0.6B
EMBEDDING·32K CTX· EMBEDDING· MULTILINGUAL
321
TOK/S · 11% VRAM
›
S
Falcon-H1R Tiny 0.6B0.6B
FALCON·32K CTX· CHAT· REASONING
321
TOK/S · 11% VRAM
›
S
Falcon Perception 0.6B0.6B
FALCON·4K CTX· VISION
321
TOK/S · 11% VRAM
›
S
Qwen 1.5 0.5B0.62B
QWEN·32K CTX· CHAT
311
TOK/S · 11% VRAM
›
S
LFM2 700M0.74B
LFM·125K CTX· CHAT· TOOL_USE
261
TOK/S · 12% VRAM
›
S
Qwen3 0.6B0.75B
QWEN·32K CTX· CHAT· REASONING
257
TOK/S · 12% VRAM
›
S
GPT-2 Large 774M0.81B
GPT2·1K CTX· CHAT
238
TOK/S · 12% VRAM
›
S
Qwen 3.5 0.8B0.87B
QWEN·256K CTX· CHAT· CODING· MULTILINGUAL
222
TOK/S · 13% VRAM
›
S
InternVL3 1B0.94B
OTHER·32K CTX· CHAT· VISION
205
TOK/S · 13% VRAM
›
S
MiniCPM5 1B1.08B
MINICPM·128K CTX· CHAT· REASONING· MULTILINGUAL
179
TOK/S · 14% VRAM
›
S
TinyLlama 1.1B1.1B
LLAMA·2K CTX· CHAT
175
TOK/S · 15% VRAM
›
S
LFM2.5-1.2B-Thinking1.2B
LFM·122K CTX· CHAT· REASONING· TOOL_USE
161
TOK/S · 15% VRAM
›
S
Llama-3.2-1B1.2B
LLAMA·4K CTX· CHAT
161
TOK/S · 15% VRAM
›
S
LFM2 1.2B1.2B
LFM·125K CTX· CHAT· TOOL_USE· MULTILINGUAL
161
TOK/S · 15% VRAM
›
S
Zamba2 1.2B1.2B
OTHER·4K CTX· CHAT
161
TOK/S · 15% VRAM
›
S
EXAONE-4.0-1.2B1.3B
EXAONE·64K CTX· CHAT
148
TOK/S · 16% VRAM
›
S
OPT 1.3B1.3B
OPT·2K CTX· CHAT
148
TOK/S · 16% VRAM
›
S
MiniCPM-V 4.61.3B
OTHER·256K CTX· CHAT· VISION
148
TOK/S · 16% VRAM
›
S
DeepSeek Coder 1.3B1.35B
DEEPSEEK·16K CTX· CODING
143
TOK/S · 16% VRAM
›
S
Phi-1 1.3B1.42B
PHI·2K CTX· CODING
136
TOK/S · 17% VRAM
›
S
Phi-1.5 1.3B1.42B
PHI·2K CTX· CHAT· CODING
136
TOK/S · 17% VRAM
›
S
LFM2 8B A1B8.3BMoE
LFM·125K CTX· CHAT· TOOL_USE· MULTILINGUAL
129
TOK/S · 70% VRAM
›
▸ NEXT STEP

Get personalized recommendations.

See ranked models with benchmark scores, run commands, and precise speed estimates for your Radeon Pro 580.

WHAT THIS CARD IS WORTH

Radeon Pro 580 holds 201 of the models in our catalogue and is, in practice, a IQ4_XS card — the largest it takes is Falcon3-10B 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
Falcon3-10B10.3B · IQ4_XS7.1 GB20 tok/sEST
Bamba 9B v29.78B · Q4_K_M7.2 GB19 tok/sEST
RecurrentGemma 9B9.63B · Q4_K_M6.8 GB20 tok/sEST
glm-4-9b9.4B · Q4_K_M6.6 GB20 tok/sEST
Yi 1.5 9B9B · Q4_K_M6.5 GB21 tok/sEST
Yi Coder 9B9B · Q4_K_M6.5 GB21 tok/sEST
Ministral 3 8B8.92B · Q4_K_M6.6 GB21 tok/sEST
Ministral 3 8B Reasoning8.92B · Q4_K_M6.6 GB21 tok/sEST
AGAINST CARDS OF SIMILAR MEMORY
M3 Pro (18GB)12 GB$1,5991.5 tok/s per $100
M1 Pro (16GB)11 GB$9992.9 tok/s per $100
M2 Pro (16GB)11 GB$1,2992.2 tok/s per $100
M4 (16GB)11 GB$4994.0 tok/s per $100
Radeon Pro 5808 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 580 — 8 GB VRAM.

▸ RADEON PRO 580 SPEC
BRAND
AMD
VRAM
8 GB GDDR5
BANDWIDTH
217 GB/s
FP16 COMPUTE
5.5 TFLOPS
FP32 COMPUTE
5.5 TFLOPS
STREAM PROCESSORS
2,304
TDP
185 W
ARCHITECTURE
GCN 4.0
▸ AI CAPABILITY
201/ 449 models @ Q4

With 8 GB VRAM and 217 GB/s bandwidth, this GPU handles models up to 9.63B parameters.

Speed ≈ bandwidth / model_size × efficiency. A 7B model at Q4 runs at ~25 tok/s.

RENT IT — LIVE PRICES
Checking what the clouds are charging…
§ 01TOP MODELS FOR RADEON PRO 580
201 FIT · SHOWING 20
MODELSIZEVRAM Q4TOK/SAVG
RecurrentGemma 9B9.63B6.4 GB2035.0
glm-4-9b9.4B6.2 GB2120.5
gemma-2-9b9.2B6.1 GB2130.2
Yi 1.5 9B9B6.0 GB2130.3
Yi Coder 9B9B6.0 GB2135.8
Ministral 3 8B8.92B5.9 GB2225.7
Ministral 3 8B Reasoning8.92B5.9 GB22—
NVIDIA-Nemotron-Nano-9B-v28.9B5.9 GB2244.2
InternLM3 8B Instruct8.8B5.9 GB2238.7
Gemma 1 7B8.54B5.7 GB2324.7
CodeGemma 7B8.54B5.7 GB2340.2
LFM2 8B A1B8.3B5.6 GB12924.3
Seed-Coder 8B Instruct8.25B5.5 GB2334.1
Seed-Coder 8B Reasoning8.25B5.5 GB2332.9
DeepSeek R1-0528 Qwen3 8B8.2B5.5 GB2436.3
Qwen3-8B8.2B5.5 GB2443.3
Granite 3.0 8B8.17B5.5 GB2436.4
Granite 3.1 8B8.17B5.5 GB2438.6
Command-R7B8.03B5.4 GB2435.3
Aya Expanse 8B8B5.4 GB2427.8