⚡ Arm Pulse
v1.0.0
KleidiAI-powered LLM Inference Benchmark Suite + MCP Server for Arm64
🖥️ Platform:
Loading...
Best Tokens/sec
—
Run benchmark to see
Avg Time to First Token
—
milliseconds
Speedup Factor
—
Q4 vs Q8
Models Tested
—
quantization levels
📊 Benchmark Results
Model
Quant
Tokens/sec
TTFT (ms)
RAM (MB)
Size (GB)
Speed
Click "Run Benchmark" to populate results
🧪 Try Inference
What is KleidiAI and how does it optimize AI on Arm processors?
Llama-3.2-3B Q4_K_M (Fast)
Llama-3.2-3B Q8_0 (Quality)
▶ Generate
📊 Run Benchmark
Output will appear here...