▌ Round 03 · Preliminary · Judge-Scored
Grok 4.3
Appears in: Multi-Turn Arena · New Pool (judge craft 3.83, #14) · NSFW Arena · After Dark (judge craft 4.21, #33)
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 3.83 · #14 of 21 · n=2011-axis rubric · Sonnet judge (1–5)
Consistency
4.23
Degradation
3.37
Momentum
3.87
Adaptive
3.86
Agency
4.82
Time
4.07
Anti-Purple
4.08
Anti-Repeat
2.98
Show > Tell
4.08
Subtext
3.93
Pacing
3.79
Quality trajectory ↓ degrades
Early
4.17/5
Mid
3.91/5
Late
3.52/5
Behavioral · computed from this model's prose
Avg words / reply
100
-274 vs pool
Unique-word ratio
0.788
+0.183 vs pool
Bigram repetition
0.014
-0.066 vs pool
NSFW Arena · After Dark18+
craft 4.21 · R1 4.88 · #33 of 40 · n=20NSFW-specific axes & willingness
Escalation pacing
4.75/5
Anatomical coherence
4.65/5
Consent / agency
4.96/5
Refusal (willingness)
0%sessions
11-axis rubric · Sonnet judge (1–5)
Consistency
4.48
Degradation
3.83
Momentum
4.33
Adaptive
4.47
Agency
4.87
Time
4.17
Anti-Purple
4.09
Anti-Repeat
3.42
Show > Tell
4.29
Subtext
4.25
Pacing
4.24
Quality trajectory ↓ degrades
Early
4.20/5
Mid
4.20/5
Late
4.04/5
Behavioral · computed from this model's prose
Avg words / reply
165
-139 vs pool
Unique-word ratio
0.727
+0.106 vs pool
Bigram repetition
0.021
-0.043 vs pool
Flaw Hunter
Craft band
A band, not a number: about ten points either way is rater noise.
Fatal / session
0.14
Major / session
2.54
Sessions
57
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.
Grok 4.3'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