AMD RX 9060 XT (16GB)

The AMD RX 9060 XT (16GB) provides 16 GB of memory with 320 GB/s bandwidth, supporting local inference via ROCm and Vulkan acceleration.

Memory

16 GB

Usable Memory Ceiling

16 GB

Bandwidth

320 GB/s

Max Inference Class

14B Models (Q4)

TDP

180W

VRAM Scaling by Context Length

Estimated memory requirements as context length increases. Curves crossing above the reference line indicate Out of Memory (OOM).

SmolVLM2 256M Video Instruct (0.26B)SmolLM3 3B (3B)Falcon3-10B (10B)Mixtral-8x7B-v0.1 (46.7B (7B active))

Precision:

Best Models to Run on AMD RX 9060 XT (16GB)

Inference throughput, memory compatibility, and generation speeds for open-weights LLMs on AMD RX 9060 XT (16GB).

Sort by:

Context:

KV Cache:

GPUs:

RankModelParametersPrecisionTPS (Approx.)TTFTCan Run

#51

Inkling 2 Small

2B (17B active)

INT4
Q8
FP16

26 tok/s

15 tok/s

8 tok/s

~1.7s

~1.8s

~1.8s

#55

Qwen 3.8 27B

27B

INT4

15 tok/s

~3.3s

#99

Qwen3.5-27B

27B

INT4

15 tok/s

~3.3s

#115

Phi-4 Reasoning Plus

14B

INT4

22 tok/s

~1.5s

#116

MiMo V2 Flash

15B (309B active)

INT4

1 tok/s

~33.7s

#122

Gemma 4 26B A4B

25.2B (3.8B active)

INT4

70 tok/s

~558 ms

#129

Qwen3.5-9B

9B

INT4
Q8

44 tok/s

26 tok/s

~973 ms

~988 ms

#141

Gemma 4 12B

11.95B

INT4
Q8

25 tok/s

15 tok/s

~1.3s

~1.5s

#145

Qwen3.5-4B

4B

INT4
Q8
FP16

81 tok/s

53 tok/s

30 tok/s

~450 ms

~457 ms

~472 ms

#150

GPT-OSS 20B

21B (3.6B active)

INT4

31 tok/s

~582 ms

#153

K2 Horizon 7B

7B

INT4
Q8
FP16

54 tok/s

33 tok/s

16 tok/s

~763 ms

~775 ms

~936 ms

#157

NVIDIA Nemotron 3 Nano 30B-A3B

3.5B (30B active)

INT4
Q8
FP16

14 tok/s

8 tok/s

4 tok/s

~3.1s

~3.1s

~3.2s

#164

Granite 4.2 8B

8B

INT4
Q8

48 tok/s

29 tok/s

~867 ms

~880 ms

#167

Gemma 4 E4B

8B

INT4
Q8

48 tok/s

29 tok/s

~867 ms

~880 ms

#172

Phi-4

14B

INT4
Q8

30 tok/s

16 tok/s

~1.5s

~1.8s

#173

Gemma 3 27B

27B

INT4

15 tok/s

~3.3s

#185

DeepSeek-R1 14B

14B

INT4
Q8

31 tok/s

16 tok/s

~1.5s

~1.8s

#189

Ministral 3 14B

14B

INT4

22 tok/s

~1.5s

#192

Gemma 3 12B

12B

INT4
Q8

34 tok/s

19 tok/s

~1.3s

~1.4s

#194

Mistral-Small-2501

24B

INT4

18 tok/s

~2.8s

Inference Capacity

Maximum model parameter class runnable by weight precision.

Context:

4-bit

Q4

14B Models

8-bit

Q8

8B Models

16-bit

FP16

3B Models

Fine-Tuning Capacity

Hardware capacity for local training and adapter fine-tuning.

Context Length:

QLoRA

4-bit

14B Models

LoRA

16-bit

8B Models

Full Parameter Training

1B Models

Workload Recommendations

Guidance for optimal precision and operational ceilings on AMD RX 9060 XT (16GB).

Inference Sweet Spot

Optimal for 8B models across extended context (32k+). 14B models fit comfortably at Q4 precision with standard context.

Bandwidth & Speed Profile

With 320 GB/s aggregate bandwidth, batch size 1 inference operates in a memory-bandwidth bound regime. Generates approximately 71 tok/s on an 8B Q4 model and 8 tok/s on a 70B Q4 model.

Frequently Asked Questions