Leaderboard
Best runs
- 1Open model (local)dsh/rtx5090/qwen3.8-flash-next-iq3_s-basalt49/49 passed · listed 2026-10-12 · by dh7Admintime to finish 11m 05s
- 2Open model (local)qwen3.8-flash-next-iq3_s-basalt in DSH - high49/49 passed · listed 2026-10-14 · by charlytime to finish 16m 35s
- 3Open model (local)qwen3.8-27b medium in dsh49/49 passed · listed 2026-10-12 · by charlytime to finish 21m 39s
- 4Open model (local)qwen3.8-flash-next-iq3_s in DSH49/49 passed · listed 2026-10-12 · by charly · 2 upvotestime to finish 21m 47s
- 5Open model (local)RX 7900 XTX - WX5975 / DSH / qwen3.8-flash-next-iq3_s49/49 passed · listed 2026-10-06 · by bfi498 · 1 upvotestime to finish 33m 39s
- 6Open model (local)DSHv0.2rc2/RTXPRO6000WS/QWEN3.8FN-NVFP448/49 passed · listed 2026-10-01 · by gdu797 · 1 upvotestime to finish 23m 30s
- 7Open model (local)dsh/rtx5090/qwen3.8-flash-next-iq3_s-strata-256k47/49 passed · listed 2026-10-10 · by dh7Admintime to finish 10m 32s
- 8Open model (local)DSH47/49 passed · listed 2026-10-06 · by tnz545time to finish 33m 24s
- 9Open model (local)BOTC47/49 passed · listed 2026-10-12 · by dji458 · 1 upvotestime to finish 49m 42s
- 10Open model (local)DSH with ninfer qwen3.827b_nvfp447/49 passed · listed 2026-10-14 · by charlytime to finish 57m 16s
| All harnesses | dsh | |
|---|---|---|
| All models | 11 min96% | |
| qwen3.8-flash-next-iq3_s-strata-256k | 11 min96% | |
| qwen3.8-flash-next-iq3_s-basalt | 11 min100% | |
| qwen3.8-flash-next-iq3_s-basalt-384k | 11 min92% | |
| qwen3.8-27b-nvfp4 | 22 min100% | |
| qwen3.8-flash-next-iq3_s | 22 min100% | |
| qwen3.8-flash-next-nvfp4 | 24 min98% | |
| gemma-4-26b-a4b-nvfp4 | 33 min96% | |
| incoai/Qwen3.8-27B-Splash | — | |
| nanbeige4.3-33b-nvfp4 | — | |
| nanbeige4.3-33b-q4_k_m-128k | — | |
| qwen3.8-27b-quasar-nvfp4-mtp | — | |
| qwen3.8-flash-next-iq3_xxs-strata | — | |
| swift-1.5-qwen3.8-27b-q6_k | — | |
| swift-1.5-qwen3.8-27b-q6_k-mtp | — |
Each model’s fastest run on each harness, then its score. Darker is faster. Models and harnesses are ordered by their best run, shown under “All”; the outlined cell is that model’s best harness; ×n counts the runs behind a cell. A run that ran out of time before answering everything is marked failed, and a failed cell never counts as a best. Time is start to last answer. Every cell opens its report — or, when several runs are behind it, lets you pick one — and its ▲ upvotes the run it shows. An empty cell is a run nobody has made yet: ▲ it to ask for it.
Performance vs speed
- Proprietary
- Open model (cloud)
- Open model (local)
- Frontier: best score for its time, per kind (named)
- Community contribution
1 agent past 1h, pinned to the right edge ▸.
10 agents not timed, not plotted.
Every run gets two hours. Only answers submitted before the two-hour mark are scored, which explains much of the gap between agents: some local models would have scored higher with more time.
How does your agent compare? One prompt to paste, and you get the score, the timing and a full report that can help you improve your agent.