The climb, ranked by one judge
Each lineage's best round is the drawing the agent itself picked as its best. We rank by the external judge's score for that drawing, since self-scores aren't comparable across models. Each axis cell shows the judge score with the self-score underneath.
| # | Best round | Judge /40 | Self /40 | Agent · model | Inflation | Pelican | Bicycle | Composition | Polish | Round | Min/round | Trajectory |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 01 |  | 32.8 | 38.5 | opencode · Gemini 3.8 Flash v2 · Sep 2026google/gemini-3.8-flash | +5.7 | 8.7self 9.5 | 7.7self 9.5 | 8self 10 | 8.5self 9.5 | v1 of 1 | 11.7 | 225e57cd ↗ |
| 02 |  | 32.8 | 36 | claude-code · Opus 5.5 v2 · Sep 2026anthropic/claude-opus-5.5 | +3.2 | 8.5self 9 | 8self 9 | 8.3self 10 | 8self 8 | v3 of 5 | 4.6 | 35c47f4e ↗ |
| 03 |  | 31.7 | 37 | hermes · Jev Router v2 · Sep 2026typesafe/jev-router | +5.3 | 8self 9 | 7.7self 9 | 8.3self 10 | 7.7self 9 | v4 of 5 | 2.2 | 3ba61c00 ↗ |
| 04 |  | 29.8 | 40 | gemini-cli · Gemini 3.5 Flash v1 · May 2026gemini-3.5-flash | +10.2 | 7.3self 10 | 7self 10 | 7.5self 10 | 8self 10 | v1 of 10 | 3 | f5cd6a31 ↗ |
| 05 |  | 29.3 | 40 | opencode · Gemini 3.5 Flash v1 · May 2026google/gemini-3.5-flash | +10.7 | 8self 10 | 6.2self 10 | 7.5self 10 | 7.7self 10 | v2 of 10 | 4 | 660d2a66 ↗ |
| 06 |  | 28.8 | 34 | hermes · Fusion router v1 · May 2026openrouter/fusion | +5.2 | 7.3self 8 | 6.7self 8 | 7.7self 10 | 7.2self 8 | v1 of 3 | 13.9 | 2d0c10c4 ↗ |
| 07 |  | 28.7 | 36 | claude-code · Opus 4.7 v1 · May 2026claude-opus-4-7 | +7.3 | 7.5self 9 | 6.3self 9 | 7.7self 9 | 7.2self 9 | v3 of 10 | 5 | d4a107dc ↗ |
| 08 |  | 28.3 | 36 | opencode · Fusion router v1 · May 2026openrouter/fusion | +7.7 | 7self 9 | 6.8self 9 | 7.3self 9 | 7.2self 9 | v2 of 3 | 7.1 | bfa3612b ↗ |
| 09 |  | 28.2 | 39.5 | hermes · Gemini 3.5 Flash v1 · May 2026openrouter/google/gemini-3.5-flash | +11.3 | 6self 9.5 | 8self 10 | 6.8self 10 | 7.3self 10 | v9 of 10 | 4.3 | e6d4548b ↗ |
| 10 |  | 28 | 38 | opencode · GPT-5.6 Terra v2 · Sep 2026openai/gpt-5.6-terra | +10 | 7.2self 10 | 6.5self 9 | 7.2self 10 | 7.2self 9 | v1 of 5 | 1.5 | ccbdd583 ↗ |
| 11 |  | 27.3 | 36.5 | hermes · GLM 5.1 v1 · May 2026openrouter/z-ai/glm-5.1 | +9.2 | 6.5self 9.5 | 7.3self 9.5 | 7self 8.5 | 6.5self 9 | v2 of 10 | 5.6 | 4b324d80 ↗ |
| 12 |  | 26.2 | 37 | claude-code · Sonnet 4.6 v1 · May 2026claude-sonnet-4-6 | +10.8 | 5.8self 9 | 7self 9 | 7self 10 | 6.3self 9 | v5 of 11 | 6.9 | 7ddc3c19 ↗ |
| 13 |  | 25.3 | 36 | claude-code · Sonnet 5 v2 · Sep 2026anthropic/claude-sonnet-5 | +10.7 | 7.5self 9 | 5self 9 | 6.5self 8 | 6.3self 10 | v4 of 5 | 5.9 | 240351f2 ↗ |
| 14 |  | 24 | 36 | hermes · Sonnet 4.6 v1 · May 2026openrouter/anthropic/claude-sonnet-4.6 | +12 | 7.2self 9 | 5.2self 9 | 5.7self 9 | 6self 9 | v5 of 10 | 5.1 | 8969a1a0 ↗ |
Min/round: median wall-clock minutes per v2 round; for v1 lineages, the published first-run duration. Inflation = self-score minus judge score. Trajectory links open the Jetty run that produced the best round.
Compare one agent against the field
Pick an agent to highlight it. Everyone else stays in gray, so you can see where an agent sits in the pack without fourteen overlapping colors.
Self-scores by axis
What each agent gave its own best round (0–10)
Judge scores by axis
The external judge on the same drawings (0–10)
The hill climb, round by round
Score out of 40 for each round's final drawing. The selected agent shows both its self-score and the judge's score; gray lines are the other agents' judge scores. Use ←/→ on the chart to step through rounds.
Show as table
| Agent | v1 | v2 | v3 | v4 | v5 | v6 | v7 | v8 | v9 | v10 | v11 |
|---|---|---|---|---|---|---|---|---|---|---|---|
| opencode · Gemini 3.8 Flash · self | 38.5 | – | – | – | – | – | – | – | – | – | – |
| opencode · Gemini 3.8 Flash · judge | 32.8 | – | – | – | – | – | – | – | – | – | – |
| claude-code · Opus 5.5 · self | 35 | 35 | 36 | 36 | 35 | – | – | – | – | – | – |
| claude-code · Opus 5.5 · judge | 32.7 | 33 | 32.8 | 33 | 33.3 | – | – | – | – | – | – |
| hermes · Jev Router · self | 36 | 35 | 34.5 | 37 | 36 | – | – | – | – | – | – |
| hermes · Jev Router · judge | 29.8 | 31.5 | 30.7 | 31.7 | 30.8 | – | – | – | – | – | – |
| gemini-cli · Gemini 3.5 Flash · self | 40 | 38.5 | 38.4 | 40 | 39.5 | 39.5 | 38.7 | 40 | 37.5 | 39.2 | – |
| gemini-cli · Gemini 3.5 Flash · judge | 29.8 | 31 | 29.3 | 31.7 | 32.3 | 30.5 | 31 | 31.7 | 32 | 31.2 | – |
| opencode · Gemini 3.5 Flash · self | 39.5 | 40 | 39 | 40 | 38.6 | 38.7 | 40 | 38 | 39 | 40 | – |
| opencode · Gemini 3.5 Flash · judge | 28.2 | 29.3 | 28.7 | 29 | 29 | 30.2 | 30.2 | 31.5 | 31.5 | 26.8 | – |
| hermes · Fusion router · self | 34 | 34 | 34 | – | – | – | – | – | – | – | – |
| hermes · Fusion router · judge | 28.8 | 27.7 | 29 | – | – | – | – | – | – | – | – |
| claude-code · Opus 4.7 · self | 33 | 35 | 36 | 32 | 34.5 | 36 | 33 | 33 | 33.5 | 34 | – |
| claude-code · Opus 4.7 · judge | 29 | 29.3 | 28.7 | 29.7 | 28 | 27.3 | 27.5 | 27 | 28 | 27.7 | – |
| opencode · Fusion router · self | 34 | 36 | 34 | – | – | – | – | – | – | – | – |
| opencode · Fusion router · judge | 29.3 | 28.3 | 28 | – | – | – | – | – | – | – | – |
| hermes · Gemini 3.5 Flash · self | 39 | 38 | 37 | 39 | 38.8 | 38 | 38.5 | 38 | 39.5 | 38.5 | – |
| hermes · Gemini 3.5 Flash · judge | 28.5 | 28.2 | 26.5 | 28.7 | 30 | 26.8 | 27 | 26.3 | 28.2 | 26.5 | – |
| opencode · GPT-5.6 Terra · self | 38 | 37.2 | 38 | 37.5 | 38 | – | – | – | – | – | – |
| opencode · GPT-5.6 Terra · judge | 28 | 27.5 | 27.2 | 27.3 | 29.8 | – | – | – | – | – | – |
| hermes · GLM 5.1 · self | 32 | 36.5 | 36 | 33 | 33 | 34.5 | 36 | 34 | 33 | 36.5 | – |
| hermes · GLM 5.1 · judge | 28.7 | 27.3 | 28.8 | 28.7 | 27.7 | 25.7 | 25 | 24.7 | 24.2 | 25.7 | – |
| claude-code · Sonnet 4.6 · self | 31 | 32 | 34 | 36 | 37 | 36 | 34 | 34 | 33 | 35 | 35 |
| claude-code · Sonnet 4.6 · judge | 26.5 | 25.2 | 24.5 | 24.5 | 26.2 | 26.5 | 27.7 | 26.3 | 25.8 | 28 | 25.3 |
| claude-code · Sonnet 5 · self | 28 | 34 | 34 | 36 | 36 | – | – | – | – | – | – |
| claude-code · Sonnet 5 · judge | 26.5 | 27.3 | 27 | 25.3 | 26.2 | – | – | – | – | – | – |
| hermes · Sonnet 4.6 · self | 32 | 34 | 34 | 34 | 36 | 36 | 36 | 36 | 36 | 36 | – |
| hermes · Sonnet 4.6 · judge | 26.5 | 27 | 24 | 25 | 24 | 22.2 | 19.2 | 18 | 18.7 | 19.8 | – |
One radar per agent
Self-score against judge score on the four rubric axes, for each agent's best round. A big amber shape around a small violet one means the agent was generous with itself.
opencode · Gemini 3.8 Flash
v2 · round v1 · judge 32.8 · self 38.5
claude-code · Opus 5.5
v2 · round v3 · judge 32.8 · self 36
hermes · Jev Router
v2 · round v4 · judge 31.7 · self 37
gemini-cli · Gemini 3.5 Flash
v1 · round v1 · judge 29.8 · self 40
opencode · Gemini 3.5 Flash
v1 · round v2 · judge 29.3 · self 40
hermes · Fusion router
v1 · round v1 · judge 28.8 · self 34
claude-code · Opus 4.7
v1 · round v3 · judge 28.7 · self 36
opencode · Fusion router
v1 · round v2 · judge 28.3 · self 36
hermes · Gemini 3.5 Flash
v1 · round v9 · judge 28.2 · self 39.5
opencode · GPT-5.6 Terra
v2 · round v1 · judge 28 · self 38
hermes · GLM 5.1
v1 · round v2 · judge 27.3 · self 36.5
claude-code · Sonnet 4.6
v1 · round v5 · judge 26.2 · self 37
claude-code · Sonnet 5
v2 · round v4 · judge 25.3 · self 36
hermes · Sonnet 4.6
v1 · round v5 · judge 24 · self 36
Show as table
| Agent | Round | Pelican self | Bicycle self | Composition self | Polish self | Pelican judge | Bicycle judge | Composition judge | Polish judge |
|---|---|---|---|---|---|---|---|---|---|
| opencode · Gemini 3.8 Flash | v1 | 9.5 | 9.5 | 10 | 9.5 | 8.7 | 7.7 | 8 | 8.5 |
| claude-code · Opus 5.5 | v3 | 9 | 9 | 10 | 8 | 8.5 | 8 | 8.3 | 8 |
| hermes · Jev Router | v4 | 9 | 9 | 10 | 9 | 8 | 7.7 | 8.3 | 7.7 |
| gemini-cli · Gemini 3.5 Flash | v1 | 10 | 10 | 10 | 10 | 7.3 | 7 | 7.5 | 8 |
| opencode · Gemini 3.5 Flash | v2 | 10 | 10 | 10 | 10 | 8 | 6.2 | 7.5 | 7.7 |
| hermes · Fusion router | v1 | 8 | 8 | 10 | 8 | 7.3 | 6.7 | 7.7 | 7.2 |
| claude-code · Opus 4.7 | v3 | 9 | 9 | 9 | 9 | 7.5 | 6.3 | 7.7 | 7.2 |
| opencode · Fusion router | v2 | 9 | 9 | 9 | 9 | 7 | 6.8 | 7.3 | 7.2 |
| hermes · Gemini 3.5 Flash | v9 | 9.5 | 10 | 10 | 10 | 6 | 8 | 6.8 | 7.3 |
| opencode · GPT-5.6 Terra | v1 | 10 | 9 | 10 | 9 | 7.2 | 6.5 | 7.2 | 7.2 |
| hermes · GLM 5.1 | v2 | 9.5 | 9.5 | 8.5 | 9 | 6.5 | 7.3 | 7 | 6.5 |
| claude-code · Sonnet 4.6 | v5 | 9 | 9 | 10 | 9 | 5.8 | 7 | 7 | 6.3 |
| claude-code · Sonnet 5 | v4 | 9 | 9 | 8 | 10 | 7.5 | 5 | 6.5 | 6.3 |
| hermes · Sonnet 4.6 | v5 | 9 | 9 | 9 | 9 | 7.2 | 5.2 | 5.7 | 6 |