02 — Runs · v2 · jettygrowthteam · auto-evolved runbook (weakest-axis targeting)
opencode
mistralai/mistral-large-4-0
opencode driving mistralai/mistral-large-4-0 through OpenRouter, 5 hill-climb rounds from the v1 runbook.
Self-score vs judge, by axis
Best round v2, 0–10 per axis
Judge: The pelican reads clearly, with a big orange throat pouch, white body, a wing reaching forward to the handlebars and orange legs with feet on the pedals; the neck is a thin, stiff stick and the head is small. The bicycle has spoked wheels, a red frame, handlebars and a crank, but there is no visible seat and the geometry is odd: the seat stay seems to run to the rear hub, and the down tube and pedal placement are confusing. The flat-shaded style is clean, with speed lines and a ground shadow.
Score by round
Out of 40. Each round starts from the previous round's best SVG. Dashed line: the 28/40 pass threshold.
Failed attempts (retried): 9553a199.
5 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.
Scores per round
| Round | Pelican | Bicycle | Composition | Polish | Self total | Judge total | Verdict | Judge spread | Cost | Minutes | Trajectory |
|---|---|---|---|---|---|---|---|---|---|---|---|
| v1 | 7.5self 9 | 6.2self 9 | 7self 8 | 7self 8 | 34 | 27.7 | fail | 0.5 | not metered | 4.7 | 508c1c1a |
| v2 | 7.7self 9 | 6.3self 9 | 7.3self 9 | 7.5self 9 | 36 | 28.8 | pass | 0.5 | not metered | 5.4 | cec17694 |
| v3 | 7.8self 9 | 6.3self 8 | 7.7self 9 | 7.3self 8 | 34 | 29.2 | pass | 1.5 | not metered | 5.3 | c8da2786 |
| v4 | 7.8self 9 | 7.3self 9 | 7.5self 9 | 7.7self 8 | 35 | 30.3 | pass | 1.5 | not metered | 5.8 | c52d8e82 |
| v5 | 7.5self 9 | 7.3self 9 | 7.7self 9 | 7.5self 9 | 36 | 30 | pass | 1 | not metered | 7.8 | 0bcb4a6c |