AMD/ CDNA 4.0
AMD Radeon Instinct MI350X
288 GB
VRAM (Flagship)
8190 GB/s
Memory Bandwidth
284 / 290
Models fit at Q4 (98%)
~1040 tok/s
7B Q4 speed (Blazing)
Model coverage at Q498% of all models
Specifications
VRAM
288 GB
Bandwidth
8190 GB/s
FP16 Compute
576.7 TFLOPS
TDP
1000W
Memory
HBM3e
Architecture
CDNA 4.0
Stream Processors
16,384
Compute Units
128
PCIe
Gen 5 x16
282
Fast models (>30 tok/s)
Real-time chat speed
284
Usable models (>10 tok/s)
Comfortable for all tasks
284
Total compatible
Fit in VRAM at Q4
Compatible Models(284)
S
GPT-2 124M0.124B
gpt21K ctxchat
6.5
quality
VRAM0%
0.4 GB
58710
tok/s
58710
tok/s
0% VRAM
S
SmolLM2 135M0.135B
smollm2K ctxchat
7
quality
VRAM0%
0.4 GB
53926
tok/s
53926
tok/s
0% VRAM
S
Nomic Embed Text v1.50.137B
embedding8K ctxembedding
VRAM0%
0.3 GB
53139
tok/s
53139
tok/s
0% VRAM
S
mxbai-embed-large-v10.335B
embedding1K ctxembedding
VRAM0%
0.4 GB
21731
tok/s
21731
tok/s
0% VRAM
S
Snowflake Arctic Embed L0.335B
embedding1K ctxembedding
VRAM0%
0.4 GB
21731
tok/s
21731
tok/s
0% VRAM
S
GPT-2 Medium 345M0.345B
gpt21K ctxchat
5.9
quality
VRAM0%
0.5 GB
21101
tok/s
21101
tok/s
0% VRAM
S
SmolLM2 360M0.36B
smollm8K ctxchat
8.2
quality
VRAM0%
0.5 GB
20222
tok/s
20222
tok/s
0% VRAM
S
Qwen 1.5 0.5B0.5B
qwen32K ctxchat
9.9
quality
VRAM0%
0.8 GB
14560
tok/s
14560
tok/s
0% VRAM
S
Qwen 2.5 0.5B0.5B
qwen32K ctxchat
19.4
quality
VRAM0%
0.8 GB
14560
tok/s
14560
tok/s
0% VRAM
S
BGE-M30.568B
embedding8K ctxembedding
VRAM0%
0.6 GB
12817
tok/s
12817
tok/s
0% VRAM
S
Qwen3 0.6B0.6B
qwen32K ctxchatreasoning
20.2
quality
VRAM0%
0.6 GB
12133
tok/s
12133
tok/s
0% VRAM
S
GPT-2 Large 774M0.774B
gpt21K ctxchat
5.6
quality
VRAM0%
0.7 GB
9406
tok/s
9406
tok/s
0% VRAM
S
Qwen3-0.6B0.8B
qwen40K ctxchat
12.3
quality
VRAM0%
0.9 GB
9100
tok/s
9100
tok/s
0% VRAM
S
Qwen3.5-0.8B0.9B
qwen256K ctxchat
20.5
quality
VRAM0%
1.0 GB
8089
tok/s
8089
tok/s
0% VRAM
S
Falcon3-1B1B
falcon32K ctxchat
42
quality
VRAM0%
1.0 GB
7280
tok/s
7280
tok/s
0% VRAM
S
InternLM2 1B1B
internlm32K ctxchat
VRAM0%
1.0 GB
7280
tok/s
7280
tok/s
0% VRAM
S
TinyLlama-1.1B1.1B
llama2K ctxchat
13.6
quality
VRAM0%
1.1 GB
6618
tok/s
6618
tok/s
0% VRAM
S
TinyLlama 1.1B1.1B
llama2K ctxchat
13.6
quality
VRAM0%
1.1 GB
6618
tok/s
6618
tok/s
0% VRAM
S
Llama-3.2-1B1.2B
llama4K ctxchat
13.6
quality
VRAM0%
1.2 GB
6067
tok/s
6067
tok/s
0% VRAM
S
LFM2.5-1.2B-Thinking1.2B
lfm122K ctxchatreasoningtool_use
18.3
quality
VRAM0%
0.7 GB
6067
tok/s
6067
tok/s
0% VRAM
S
DeepSeek Coder 1.3B1.3B
deepseek16K ctxcoding
16.8
quality
VRAM0%
1.2 GB
5600
tok/s
5600
tok/s
0% VRAM
S
OPT 1.3B1.3B
opt2K ctxchat
5.3
quality
VRAM0%
1.2 GB
5600
tok/s
5600
tok/s
0% VRAM
S
Phi-1 1.3B1.3B
phi2K ctxcoding
7.2
quality
VRAM0%
1.2 GB
5600
tok/s
5600
tok/s
0% VRAM
S
Phi-1.5 1.3B1.3B
phi2K ctxchatcoding
7.2
quality
VRAM0%
1.2 GB
5600
tok/s
5600
tok/s
0% VRAM
S
GPT-2 XL 1.5B1.5B
gpt21K ctxchat
5.1
quality
VRAM0%
1.3 GB
4853
tok/s
4853
tok/s
0% VRAM
S
Qwen2.5-Coder-1.5B1.5B
qwen32K ctxcodingchat
22.8
quality
VRAM0%
1.3 GB
4853
tok/s
4853
tok/s
0% VRAM
S
Qwen2 Math 1.5B1.5B
qwen4K ctxreasoning
22.8
quality
VRAM0%
1.3 GB
4853
tok/s
4853
tok/s
0% VRAM
S
Qwen 2.5 1.5B1.5B
qwen32K ctxchatcoding
30.2
quality
VRAM0%
1.3 GB
4853
tok/s
4853
tok/s
0% VRAM
S
Yi Coder 1.5B1.5B
yi125K ctxcoding
14.6
quality
VRAM0%
1.3 GB
4853
tok/s
4853
tok/s
0% VRAM
S
stablelm-2-1_6b1.6B
stablelm4K ctxchat
9.5
quality
VRAM0%
1.4 GB
4550
tok/s
4550
tok/s
0% VRAM
S
Qwen3 1.7B1.7B
qwen32K ctxchatreasoning
32.6
quality
VRAM0%
1.4 GB
4282
tok/s
4282
tok/s
0% VRAM
S
SmolLM2 1.7B1.71B
smollm8K ctxchat
16.2
quality
VRAM0%
1.4 GB
4257
tok/s
4257
tok/s
0% VRAM
S
Qwen 1.5 1.8B1.8B
qwen32K ctxchat
22.8
quality
VRAM1%
1.5 GB
4044
tok/s
4044
tok/s
1% VRAM
S
Moondream2 1.9B1.9B
other2K ctxvisionchat
26.2
quality
VRAM0%
1.4 GB
3832
tok/s
3832
tok/s
0% VRAM
S
Granite 4.0 Tiny7BMoE
granite125K ctxchatcodingmultilingual
49.2
quality
VRAM2%
4.4 GB
3640
tok/s
3640
tok/s
2% VRAM
S
Gemma 1 2B2B
gemma8K ctxchat
20.2
quality
VRAM1%
1.6 GB
3640
tok/s
3640
tok/s
1% VRAM
S
Granite 3.0 2B2B
granite128K ctxchatcoding
35.8
quality
VRAM1%
1.6 GB
3640
tok/s
3640
tok/s
1% VRAM
S
Granite 3.1 2B2B
granite128K ctxchatcoding
37.8
quality
VRAM1%
1.6 GB
3640
tok/s
3640
tok/s
1% VRAM
S
Qwen2-VL 2B2.21B
qwen32K ctxchatvision
28.3
quality
VRAM1%
1.7 GB
3294
tok/s
3294
tok/s
1% VRAM
S
Qwen3.5-2B2.3B
qwen256K ctxchat
15.9
quality
VRAM1%
1.8 GB
3165
tok/s
3165
tok/s
1% VRAM
Get personalized recommendations
See ranked models with benchmark scores, run commands, and precise speed estimates for your Radeon Instinct MI350X.