bench : add --offline

[angt]

Overview

Additional information

$ ./build/bin/llama bench -hf unsloth/Qwen3.5-0.8B-GGUF
| model                          |       size |     params | backend    | threads |            test |                  t/s |
| ------------------------------ | ---------: | ---------: | ---------- | ------: | --------------: | -------------------: |
| qwen35 0.8B Q4_K - Medium      | 497.39 MiB |   752.39 M | CPU        |      16 |           pp512 |       749.59 ± 15.36 |
| qwen35 0.8B Q4_K - Medium      | 497.39 MiB |   752.39 M | CPU        |      16 |           tg128 |         86.04 ± 0.43 |

build: 70b54e140 (9608)

$ ./build/bin/llama bench -hf unsloth/Qwen3.5-0.8B-GGUF --offline
| model                          |       size |     params | backend    | threads |            test |                  t/s |
| ------------------------------ | ---------: | ---------: | ---------- | ------: | --------------: | -------------------: |
| qwen35 0.8B Q4_K - Medium      | 497.39 MiB |   752.39 M | CPU        |      16 |           pp512 |        786.55 ± 2.95 |
| qwen35 0.8B Q4_K - Medium      | 497.39 MiB |   752.39 M | CPU        |      16 |           tg128 |         85.58 ± 0.39 |

build: 70b54e140 (9608)

Requirements

• I have read and agree with the contributing guidelines
• AI usage disclosure: YES