▌ Round 03 · Preliminary · Judge-Scored
Qwen 3.7 Max
Appears in: Multi-Turn Arena · New Pool (judge craft 4.28, #10) · NSFW Arena · After Dark (judge craft 4.46, #21)
Judge-scored only: no human arena votes are included here. The human ranking can move or invert these numbers. There is no human ELO, composite, or cost/latency for this model on this page.
Multi-Turn Arena · New Pool
craft 4.28 · #10 of 21 · n=2011-axis rubric · Sonnet judge (1–5)
Consistency
4.54
Degradation
4.18
Momentum
4.42
Adaptive
4.33
Agency
4.75
Time
4.28
Anti-Purple
4.26
Anti-Repeat
3.76
Show > Tell
4.54
Subtext
4.24
Pacing
4.28
Quality trajectory ↑ builds
Early
4.16/5
Mid
4.36/5
Late
4.29/5
Behavioral · computed from this model's prose
Avg words / reply
294
-79 vs pool
Unique-word ratio
0.648
+0.043 vs pool
Bigram repetition
0.049
-0.031 vs pool
NSFW Arena · After Dark18+
craft 4.46 · R1 4.97 · #21 of 40 · n=20NSFW-specific axes & willingness
Escalation pacing
4.79/5
Anatomical coherence
4.71/5
Consent / agency
4.92/5
Refusal (willingness)
0%sessions
11-axis rubric · Sonnet judge (1–5)
Consistency
4.67
Degradation
4.34
Momentum
4.59
Adaptive
4.61
Agency
4.83
Time
4.29
Anti-Purple
4.26
Anti-Repeat
4.03
Show > Tell
4.49
Subtext
4.41
Pacing
4.42
Quality trajectory ↑ builds
Early
4.22/5
Mid
4.46/5
Late
4.45/5
Behavioral · computed from this model's prose
Avg words / reply
315
+11 vs pool
Unique-word ratio
0.626
+0.005 vs pool
Bigram repetition
0.049
-0.015 vs pool
Flaw Hunter
Craft band
A band, not a number: about ten points either way is rater noise.
Fatal / session
0.04
Major / session
3.02
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 flaw bands across new and returning models with that caveat in mind. The current model cards carry the newer single-rater band.
This is the judges' call, not the crowd's.
Qwen 3.7 Max's scores here are judge-only. Round 03's human voting has closed; its board holds the crowd's ranking.
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