Compact responses and strong performance on the player-control and perspective probes, with middling prose weakened by repetition and convenient plot turns. Its content limits were not tested in round 4.
Response coverage 99.6% · 219 of 220 turns answered.
Round 04
Willingness and judgment
No Round 04 willingness result for this model, so its content limits are untested here.
Trap probes
Failure modes
Agency violations0.0% [0.0% to 8.2%] · 43 probes
POV/tense breaks0.0% [0.0% to 10.4%] · 33 probes
Trap modes, pooled5.9% [1.6% to 19.1%] · 2/34
rank 17 of 59 models carrying all nine modes
Per trap mode, counts not rates: 2 to 9 probes each, so a percentage would not survive one probe changing.
A band, not a number, on purpose: ±10 is the rater noise floor, not a sampling error. Bands that overlap are tied, and most of the roster overlaps. The rose tick marks zero.
Sessions20
Top flawsrecycled description · purple prose · convenient world
Counted by machine
Production defects
None detected219 turns clean
Token overhead3.1x · billed per visible character, against the prose floor; not a price
Counted by machine with no judge involved: the most trustworthy block on the card.
LLM judge
Subjective band
Composite
15
The band spans where three judge families put this model. The axis is one judge's (single-judge sonnet 5); another judge shifts everyone by about a point.
Collaboration · less reliable
15
Engagement · less reliable
15
Tone consistency · less reliable
15
Text statistics, no judge
Behavioral
Avg words137 · population 320
Unique-word ratio0.796 · population 0.631
Phrase repetition0.015 · population 0.060
Round 01 arena
Community rank
Round 01 rank4 of 11 · arena ELO 1517 ± 76
Rounds 02 and 03 · set beside, never converted
Across rounds
Round 04 changed the judges, not the models: the Round 02 and 03 judge (Sonnet 4) also scored these Round 04 transcripts so that line continues, and nothing here is converted between judges or added up. Methodology · all returning models
Earlier rounds, as published
Round 02 human arena1432 · 95% interval 1349 to 1512 · 230 votes · rank 18 of 20 · voters read the same text as Round 04
Round 03 ranknot in Round 03's table
Round 04
Different judges, not comparable
Old judge (Sonnet 4)
15
on these Round 04 transcripts, a band over the core seeds, never a rank
New judge (Sonnet 5)tier B · fixed tiers A (best) to E on its own scale
Second judge (ChatGPT via Codex)2.7 (own scale, about 0.94 lower on average) · letter depends on judge: no
Jno J · no Round 04 willingness run
Round 04 transcriptsSame as Round 02 · the exact Round 02 texts, hash-checked; the old judge scored them in April, and April scores may sit a little low against September ones
Strength
Lowest phrase repetition (0.015 vs population 0.060)
Reviewed 2026-09-25. Editorial summaries of the current local benchmark cards, not new session ratings or general model safety claims. Willingness means delivery against this benchmark's product targets, including soft deflection; tested hard boundaries cover non-consent and frame-break only, and hard-boundary wording follows the first ask, as J does since 2026-09-25 (the second push is a separate, weaker card line). Craft bands overlap, and small samples limit comparisons. Token overhead is not monetary cost. Review these summaries when source results change. Source: rp-benchmark generate_profile_cards_v2.py at c418a40 · export r4c-20260925-b7ccaed. CC-BY 4.0.