AMD· CDNA 3.0

AMD Radeon Instinct MI308X

VRAM
192 GB
FLAGSHIP
BANDWIDTH
10300
GB/S
MODELS Q4
307/331
93%
7B Q4 SPEED
~1308
BLAZING
▸ MODEL COVERAGE @ Q493% OF ALL
▸ ESTIMATED SPEED· BY MODEL SIZE @ Q4

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

3B
~3052
TOK/S
7B
~1308
TOK/S
14B
~654
TOK/S
32B
~286
TOK/S
70B
~131
TOK/S
▸ SPECIFICATIONS
VRAM
192 GB
BANDWIDTH
10300 GB/s
FP16 COMPUTE
653.7 TFLOPS
TDP
750W
MEMORY
HBM3
ARCHITECTURE
CDNA 3.0
STREAM PROCESSORS
19,456
COMPUTE UNITS
152
PCIE
Gen 5 x16
MSRP
$12,000
307
FAST MODELS · >30 TOK/S
Real-time chat speed
307
USABLE · >10 TOK/S
Comfortable for all tasks
307
TOTAL COMPATIBLE
Fit in VRAM at Q4
▸ RENT IT IN THE CLOUD

Buying Radeon Instinct MI308X costs $15–$40k and isn’t practical for most teams. Spin one up by the hour instead:

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· 307
S
nomic-embed-text-v1.5 100M0.1B
EMBEDDING·8K CTX· CHAT
91556
TOK/S · 0% VRAM
S
GPT-2 124M0.124B
GPT2·1K CTX· CHAT
73835
TOK/S · 0% VRAM
S
SmolLM2 135M0.135B
SMOLLM·2K CTX· CHAT
67819
TOK/S · 0% VRAM
S
bge-large-en-v1.5 335M0.335B
EMBEDDING·1K CTX· CHAT
27330
TOK/S · 0% VRAM
S
mxbai-embed-large-v10.335B
EMBEDDING·1K CTX· EMBEDDING
27330
TOK/S · 0% VRAM
S
Snowflake Arctic Embed L0.335B
EMBEDDING·1K CTX· EMBEDDING
27330
TOK/S · 0% VRAM
S
GPT-2 Medium 345M0.345B
GPT2·1K CTX· CHAT
26538
TOK/S · 0% VRAM
S
SmolLM2 360M0.36B
SMOLLM·8K CTX· CHAT
25432
TOK/S · 0% VRAM
S
Falcon-H1 0.5B0.5B
FALCON·128K CTX· CHAT
18311
TOK/S · 0% VRAM
S
Qwen 1.5 0.5B0.5B
QWEN·32K CTX· CHAT
18311
TOK/S · 0% VRAM
S
Qwen 2.5 0.5B0.5B
QWEN·32K CTX· CHAT
18311
TOK/S · 0% VRAM
S
BGE-M30.568B
EMBEDDING·8K CTX· EMBEDDING
16119
TOK/S · 0% VRAM
S
Qwen3 0.6B0.6B
QWEN·32K CTX· CHAT· REASONING
15259
TOK/S · 0% VRAM
S
GPT-2 Large 774M0.774B
GPT2·1K CTX· CHAT
11829
TOK/S · 1% VRAM
S
Qwen 3.5 0.8B0.8B
QWEN·256K CTX· CHAT· CODING· MULTILINGUAL
11444
TOK/S · 1% VRAM
S
Qwen3.5-0.8B0.9B
QWEN·256K CTX· CHAT
10173
TOK/S · 1% VRAM
S
Falcon3-1B1B
FALCON·32K CTX· CHAT
9156
TOK/S · 1% VRAM
S
InternLM2 1B1B
INTERNLM·32K CTX· CHAT
9156
TOK/S · 1% VRAM
S
TinyLlama 1.1B1.1B
LLAMA·2K CTX· CHAT
8323
TOK/S · 1% VRAM
S
LFM2.5-1.2B-Thinking1.2B
LFM·122K CTX· CHAT· REASONING· TOOL_USE
7630
TOK/S · 1% VRAM
S
Llama-3.2-1B1.2B
LLAMA·4K CTX· CHAT
7630
TOK/S · 1% VRAM
S
DeepSeek Coder 1.3B1.3B
DEEPSEEK·16K CTX· CODING
7043
TOK/S · 1% VRAM
S
EXAONE-4.0-1.2B1.3B
EXAONE·64K CTX· CHAT
7043
TOK/S · 1% VRAM
S
OPT 1.3B1.3B
OPT·2K CTX· CHAT
7043
TOK/S · 1% VRAM
S
Phi-1 1.3B1.3B
PHI·2K CTX· CODING
7043
TOK/S · 1% VRAM
S
Phi-1.5 1.3B1.3B
PHI·2K CTX· CHAT· CODING
7043
TOK/S · 1% VRAM
S
granite-4.0-h-tiny 6.9B6.9BMoE
GRANITE·128K CTX· CHAT
6104
TOK/S · 2% VRAM
S
Falcon-H1 1.5B1.5B
FALCON·128K CTX· CHAT· CODING
6104
TOK/S · 1% VRAM
S
GPT-2 XL 1.5B1.5B
GPT2·1K CTX· CHAT
6104
TOK/S · 1% VRAM
S
Qwen2.5-Coder-1.5B1.5B
QWEN·32K CTX· CHAT· TOOL_USE· CODING
6104
TOK/S · 1% VRAM
S
Qwen2 Math 1.5B1.5B
QWEN·4K CTX· REASONING
6104
TOK/S · 1% VRAM
S
Qwen 2.5 1.5B1.5B
QWEN·32K CTX· CHAT· CODING
6104
TOK/S · 1% VRAM
S
Yi Coder 1.5B1.5B
YI·125K CTX· CODING
6104
TOK/S · 1% VRAM
S
stablelm-2-1_6b1.6B
STABLELM·4K CTX· CHAT
5722
TOK/S · 1% VRAM
S
Qwen3 1.7B1.7B
QWEN·32K CTX· CHAT· REASONING
5386
TOK/S · 1% VRAM
S
SmolLM2 1.7B1.71B
SMOLLM·8K CTX· CHAT
5354
TOK/S · 1% VRAM
S
Qwen 1.5 1.8B1.8B
QWEN·32K CTX· CHAT
5086
TOK/S · 1% VRAM
S
Moondream2 1.9B1.9B
OTHER·2K CTX· VISION· CHAT
4819
TOK/S · 1% VRAM
S
Gemma 1 2B2B
GEMMA·8K CTX· CHAT
4578
TOK/S · 1% VRAM
S
Granite 3.0 2B2B
GRANITE·128K CTX· CHAT· CODING
4578
TOK/S · 1% VRAM
▸ NEXT STEP

Get personalized recommendations.

See ranked models with benchmark scores, run commands, and precise speed estimates for your Radeon Instinct MI308X.

▸ DEVICE UNDER TEST

AMD Radeon Instinct MI308X192 GB VRAM.

RADEON INSTINCT MI308X SPEC
BRAND
AMD
VRAM
192 GB HBM3
BANDWIDTH
10300 GB/s
FP16 COMPUTE
653.7 TFLOPS
FP32 COMPUTE
81.7 TFLOPS
STREAM PROCESSORS
19,456
TDP
750 W
ARCHITECTURE
CDNA 3.0
MSRP
$12000
▸ AI CAPABILITY
307/ 331 models @ Q4

With 192 GB VRAM and 10300 GB/s bandwidth, this GPU handles models up to 236B parameters.

Speed ≈ bandwidth / model_size × efficiency. A 7B model at Q4 runs at ~1177 tok/s.

§ 01TOP MODELS FOR RADEON INSTINCT MI308X
307 FIT · SHOWING 20
MODELSIZEVRAM Q4TOK/SAVG
DeepSeek V2 236B236B144.7 GB43640.4
DeepSeek-V2.5 236B236B144.7 GB43664.0
DeepSeek-Coder-V2 236B236B144.7 GB43661.8
Qwen3-235B-A22B235.1B144.2 GB41655.9
MiniMax-M2.5 228.7B228.7B140.3 GB43662.6
Falcon 180B180B110.5 GB5151.3
bloom 176.2B176.2B108.2 GB5215.0
dots.llm1.inst 142.8B142.8B87.8 GB64
WizardLM 2 8x22B141B86.7 GB23542.4
Mixtral-8x22B140.6B86.4 GB23431.9
DBRX 132B132B81.2 GB25446.3
Qwen3.5-122B-A10B125.1B77.0 GB7345.5
Pixtral Large 124B124B76.3 GB7439.3
Mistral-Large 123B123B75.7 GB7433.5
Devstral 2 123B123B75.7 GB7438.1
Qwen 3.5 122B A10B122B75.1 GB91656.8
Nemotron 3 Super 120B120B73.8 GB76357.3
Nemotron 3 Super 120B-A12B120B73.8 GB76353.2
Mistral Small 4 119B119B73.2 GB140950.2
GPT-OSS 120B117B72.0 GB179554.1