Qwen2.5-3B-Instruct
12 records across 1 hardware.
| hardware | model | quant | backend | tok/s | pp tok/s | W | ttft s | ctx/tg | date | class | flags | source |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| MI355X | Qwen2.5-3B-Instruct | BF16 | vLLM | 2688.94 | – | – | – | 1024 | 2026-10-12 | sourced | – | source |
| MI355X | Qwen2.5-3B-Instruct | BF16 | vLLM | 2681.27 | – | – | – | 1024 | 2026-10-12 | sourced | – | source |
| MI355X | Qwen2.5-3B-Instruct | BF16 | vLLM | 953.94 | – | – | – | 8192 | 2026-10-12 | sourced | – | source |
| MI355X | Qwen2.5-3B-Instruct | BF16 | vLLM | 942.92 | – | – | – | 1024 | 2026-10-12 | sourced | – | source |
| MI355X | Qwen2.5-3B-Instruct | BF16 | vLLM | 918.01 | – | – | – | 8192 | 2026-10-12 | sourced | – | source |
| MI355X | Qwen2.5-3B-Instruct | BF16 | vLLM | 736.38 | – | – | – | 8192 | 2026-10-12 | sourced | – | source |
| MI355X | Qwen2.5-3B-Instruct | BF16 | vLLM | 570.92 | – | – | – | 8192 | 2026-10-12 | sourced | – | source |
| MI355X | Qwen2.5-3B-Instruct | BF16 | vLLM | 570.16 | – | – | – | 8192 | 2026-10-12 | sourced | – | source |
| MI355X | Qwen2.5-3B-Instruct | BF16 | vLLM | 568.81 | – | – | – | 8192 | 2026-10-12 | sourced | – | source |
| MI355X | Qwen2.5-3B-Instruct | BF16 | vLLM | 499.36 | – | – | – | 8192 | 2026-10-12 | sourced | – | source |
| MI355X | Qwen2.5-3B-Instruct | BF16 | vLLM | 443.52 | – | – | – | 8192 | 2026-10-12 | sourced | – | source |
| MI355X | Qwen2.5-3B-Instruct | BF16 | vLLM | 347.68 | – | – | – | 1024 | 2026-10-12 | sourced | – | source |