# Command: `linksite:model-bench`

**Description**: Run the multi-model refresh-content bench over pinned fixtures
**Status**: completed
**Started**: 2026-07-07 07:56:07 | **Ended**: 2026-07-07 12:39:21 | **Duration**: 4h 43m
**Jobs**: 0 dispatched / 0 completed / 0 failed

---

## Command Output

07:56:07 [INFO] → refresh × jp-1 × deepseek-v4-pro

08:02:34 [INFO] → refresh × jp-10 × deepseek-v4-pro

08:52:28 [INFO] → refresh × jp-2 × deepseek-v4-pro

08:57:32 [INFO] → refresh × jp-3 × deepseek-v4-pro

09:37:33 [INFO] → refresh × jp-4 × deepseek-v4-pro

09:53:53 [INFO] → refresh × jp-5 × deepseek-v4-pro

10:48:16 [INFO] → refresh × jp-6 × deepseek-v4-pro

11:07:19 [INFO] → refresh × jp-7 × deepseek-v4-pro

11:53:10 [INFO] → refresh × jp-8 × deepseek-v4-pro

12:03:29 [INFO] → refresh × jp-9 × deepseek-v4-pro

12:39:21 [INFO] winners → score: deepseek-v4-pro · cost: deepseek-v4-pro · speed: deepseek-v4-pro · overall: deepseek-v4-pro

12:39:21 [INFO] Artifacts: /home/guilherme/webapps/kokos-laravel2/storage/app/bench-results/6a4cb1179c7dbf42b8098f13

12:39:21 [INFO] Run id: 6a4cb1179c7dbf42b8098f13

---

## Jobs (0)

*No jobs dispatched*

*Exported at 2026-08-06T22:15:10+00:00*