| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| Qwen2.5-7B 256K context length | FastMKA | Decode Latency (ms/token)26.3 | 4 | 25d ago | |
| Qwen2.5-7B 128K context length | FastMKA | Decode Latency (ms/token)18.4 | 4 | 25d ago | |
| Qwen2.5-7B 64K context length | FastMKA | Decode Latency (ms/token)13.6 | 4 | 25d ago | |
| Qwen2.5-7B 32K context length | FastMKA | Decoding Latency (ms/token)10.3 | 4 | 25d ago | |
| Qwen2.5-7B 16K context length | FastMKA | Decode Latency (ms/token)8.4 | 4 | 25d ago | |
| Qwen2.5-7B 8K context length | FastMKA | Decode Latency (ms/token)7.1 | 4 | 25d ago | |
| Qwen2.5-7B 4K context length | FastMKA | Decode Latency (ms/token)6.2 | 4 | 25d ago |