--- title: "hermes \u00b7 Sonnet 4.6 | Pelican benchmark" url: https://evaljetty.com/pelicans/runs/agents/hermes-sonnet.html description: "hermes \u00b7 Sonnet 4.6 (openrouter/anthropic/claude-sonnet-4.6) drawing a pelican riding a bicycle: every hill-climb run with self-scores, external judge scores and Jetty trajectories." updated: 2026-09-28 publisher: Jetty (https://jetty.io) --- 02 — Runs · v1 · jettyio · auto-evolved runbook (weakest-axis targeting) # hermes openrouter/anthropic/claude-sonnet-4.6 Hermes routed to Sonnet through OpenRouter. On its first run the bird's parts floated free of the bike. Judge, best round 24/40 round v5 (agent's pick) Self-score 36/40 inflation +12 Runs 10 not recorded (v1) ![hermes · Sonnet 4.6 best round](https://evaljetty.com/pelicans/img/hermes-sonnet/v5.svg) best round v5 · 8969a1a0 ### Self-score vs judge, by axis Best round v5, 0–10 per axis agent self-scoreexternal judge **Judge:** The pelican reads well, with a large orange beak and pouch, an eye and white plumage, but its body looks upright and chicken-like and its black-tipped wings sprawl awkwardly. The bicycle has spoked wheels and a blue frame, but the frame geometry is muddled and largely hidden, the wheels crowd each other, and the feet dangle near the crank without really pedaling; the title text overlaps the wheels. ### Score by round Out of 40. Each round starts from the previous round's best SVG. Dashed line: the 28/40 pass threshold. self-scorejudgepass ≥ 28 ## 10 drawings, *oldest first.* Click a drawing to open it in the head-to-head viewer. The outlined card is the agent's best round. Verdicts are derived (code checks on the stored SVG and judge ≥ 28/40); see [Runs](https://evaljetty.com/pelicans/runs/index.html#all-runs). [![hermes · Sonnet 4.6 round v1](https://evaljetty.com/pelicans/img/hermes-sonnet/v1.svg)](https://evaljetty.com/pelicans/investigations/best-pelican.html#hermes-sonnet/v1) **v1** failjudge **26.5** · self 32 697b57d4[runbook](https://evaljetty.com/pelicans/author/versions.html#hermes-sonnet/v1) [![hermes · Sonnet 4.6 round v2](https://evaljetty.com/pelicans/img/hermes-sonnet/v2.svg)](https://evaljetty.com/pelicans/investigations/best-pelican.html#hermes-sonnet/v2) **v2** failjudge **27** · self 34 acb2ba11[runbook](https://evaljetty.com/pelicans/author/versions.html#hermes-sonnet/v2) [![hermes · Sonnet 4.6 round v3](https://evaljetty.com/pelicans/img/hermes-sonnet/v3.svg)](https://evaljetty.com/pelicans/investigations/best-pelican.html#hermes-sonnet/v3) **v3** failjudge **24** · self 34 99a389e4[runbook](https://evaljetty.com/pelicans/author/versions.html#hermes-sonnet/v3) [![hermes · Sonnet 4.6 round v4](https://evaljetty.com/pelicans/img/hermes-sonnet/v4.svg)](https://evaljetty.com/pelicans/investigations/best-pelican.html#hermes-sonnet/v4) **v4** failjudge **25** · self 34 28f94566[runbook](https://evaljetty.com/pelicans/author/versions.html#hermes-sonnet/v4) [![hermes · Sonnet 4.6 round v5](https://evaljetty.com/pelicans/img/hermes-sonnet/v5.svg)](https://evaljetty.com/pelicans/investigations/best-pelican.html#hermes-sonnet/v5) **v5** · best failjudge **24** · self 36 8969a1a0[runbook](https://evaljetty.com/pelicans/author/versions.html#hermes-sonnet/v5) [![hermes · Sonnet 4.6 round v6](https://evaljetty.com/pelicans/img/hermes-sonnet/v6.svg)](https://evaljetty.com/pelicans/investigations/best-pelican.html#hermes-sonnet/v6) **v6** failjudge **22.2** · self 36 ebc95c4b[runbook](https://evaljetty.com/pelicans/author/versions.html#hermes-sonnet/v6) [![hermes · Sonnet 4.6 round v7](https://evaljetty.com/pelicans/img/hermes-sonnet/v7.svg)](https://evaljetty.com/pelicans/investigations/best-pelican.html#hermes-sonnet/v7) **v7** failjudge **19.2** · self 36 fdf47c73[runbook](https://evaljetty.com/pelicans/author/versions.html#hermes-sonnet/v7) [![hermes · Sonnet 4.6 round v8](https://evaljetty.com/pelicans/img/hermes-sonnet/v8.svg)](https://evaljetty.com/pelicans/investigations/best-pelican.html#hermes-sonnet/v8) **v8** failjudge **18** · self 36 25323cb8[runbook](https://evaljetty.com/pelicans/author/versions.html#hermes-sonnet/v8) [![hermes · Sonnet 4.6 round v9](https://evaljetty.com/pelicans/img/hermes-sonnet/v9.svg)](https://evaljetty.com/pelicans/investigations/best-pelican.html#hermes-sonnet/v9) **v9** failjudge **18.7** · self 36 5b9cf6ae[runbook](https://evaljetty.com/pelicans/author/versions.html#hermes-sonnet/v9) [![hermes · Sonnet 4.6 round v10](https://evaljetty.com/pelicans/img/hermes-sonnet/v10.svg)](https://evaljetty.com/pelicans/investigations/best-pelican.html#hermes-sonnet/v10) **v10** failjudge **19.8** · self 36 34f10b3e[runbook](https://evaljetty.com/pelicans/author/versions.html#hermes-sonnet/v10) ### Scores per round | Round | Pelican | Bicycle | Composition | Polish | Self total | Judge total | Verdict | Judge spread | Cost | Minutes | Trajectory | | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | | v1 | 6.7self 8 | 7self 8 | 6.2self 8 | 6.7self 8 | 32 | 26.5 | fail | 1.5 | – | – | 697b57d4 | | v2 | 7.3self 8 | 6self 9 | 7.2self 9 | 6.5self 8 | 34 | 27 | fail | 1 | – | – | acb2ba11 | | v3 | 7self 9 | 5.2self 8 | 5.8self 8 | 6self 9 | 34 | 24 | fail | 1 | – | – | 99a389e4 | | v4 | 6.8self 9 | 6self 9 | 6.2self 8 | 6self 8 | 34 | 25 | fail | 1 | – | – | 28f94566 | | v5 | 7.2self 9 | 5.2self 9 | 5.7self 9 | 6self 9 | 36 | 24 | fail | 1 | – | – | 8969a1a0 | | v6 | 6.7self 9 | 5self 9 | 5.3self 9 | 5.2self 9 | 36 | 22.2 | fail | 1.5 | – | – | ebc95c4b | | v7 | 5.7self 9 | 4.3self 9 | 4.5self 9 | 4.7self 9 | 36 | 19.2 | fail | 3 | – | – | fdf47c73 | | v8 | 5.5self 9 | 4.5self 9 | 4self 9 | 4self 9 | 36 | 18 | fail | 1 | – | – | 25323cb8 | | v9 | 5.5self 9 | 4.3self 9 | 4.5self 9 | 4.3self 9 | 36 | 18.7 | fail | 1.5 | – | – | 5b9cf6ae | | v10 | 5.5self 9 | 5self 9 | 4.8self 9 | 4.5self 9 | 36 | 19.8 | fail | 0.5 | – | – | 34f10b3e | --- Source: https://evaljetty.com/pelicans/runs/agents/hermes-sonnet.html · Evals by Jetty · Run your own evals: https://jetty.io/?utm_source=evaljetty&utm_medium=llms