▌ Round 03 · Preliminary · Judge-Scored
Owl Alpha
Appears in: Multi-Turn Arena · New Pool (judge craft 4.46, #4) · NSFW Arena · After Dark (judge craft 4.59, #7)
Judge-scored only — no human arena votes yet. These numbers may move (or invert) once the human arena fills in. Pre-vote, there is no human ELO, composite, or cost/latency for this model.
Multi-Turn Arena · New Pool
craft 4.46 · #4 of 21 · n=1911-axis rubric · Sonnet judge (1–5)
Consistency
4.74
Degradation
4.45
Momentum
4.67
Adaptive
4.48
Agency
4.71
Time
4.35
Anti-Purple
4.26
Anti-Repeat
4.06
Show > Tell
4.59
Subtext
4.57
Pacing
4.38
Quality trajectory ↑ builds
Early
4.18/5
Mid
4.51/5
Late
4.52/5
Behavioral · computed from this model's prose
Avg words / reply
596
+222 vs pool
Unique-word ratio
0.533
-0.072 vs pool
Bigram repetition
0.096
+0.016 vs pool
NSFW Arena · After Dark18+
craft 4.59 · R1 4.99 · #7 of 40 · n=20NSFW-specific axes & willingness
Escalation pacing
4.85/5
Anatomical coherence
4.81/5
Consent / agency
4.93/5
Refusal (willingness)
0%sessions
11-axis rubric · Sonnet judge (1–5)
Consistency
4.80
Degradation
4.56
Momentum
4.68
Adaptive
4.63
Agency
4.89
Time
4.42
Anti-Purple
4.42
Anti-Repeat
4.27
Show > Tell
4.63
Subtext
4.70
Pacing
4.53
Quality trajectory ↑ builds
Early
4.22/5
Mid
4.51/5
Late
4.71/5
Behavioral · computed from this model's prose
Avg words / reply
451
+147 vs pool
Unique-word ratio
0.559
-0.062 vs pool
Bigram repetition
0.076
+0.012 vs pool
Flaw Hunter
Score (mean)
66.1
median 67.0
Fatal / session
0.02
Major / session
3.31
Sessions
55
Top flaws: purple prose · recycled description · narrating emotions
From the Round 03 flaw-hunter catch-up run. Scores come from two runs — compare absolute flaw scores across new vs returning models with that caveat in mind.
This is the judges' call — not the crowd's.
Owl Alpha's scores here are judge-only. Read a session it played and vote — the human ranking takes shape on the leaderboard as ballots land.
Source: rp-benchmark · scripts/build-round3-judge-preview.py. Rubric & trajectory are Claude Sonnet judge means; behavioral is computed from the model's own dialogue. No human votes included.
← Back to the Round 03 standings