-0.05
Confirmed task success · rank 25 of 46
- Unit
- IPS effect estimate, higher is better
- Range
- -0.06 to -0.03
- Sample
- 31264 observations
- Configuration
- GPT-5.6 Luna
- Measured
- 28 Sep 2026
- Not shown
- Not a success rate: the estimated effect of choosing this model as the agent's orchestrator, against an even mix of the models tested, inside Arena's own harness.
-0.04
Praise over complaint · rank 23 of 46
- Unit
- IPS effect estimate, higher is better
- Range
- -0.06 to -0.02
- Sample
- 11564 observations
- Configuration
- GPT-5.6 Luna
- Measured
- 28 Sep 2026
- Not shown
- Not a success rate: the estimated effect of choosing this model as the agent's orchestrator, against an even mix of the models tested, inside Arena's own harness.
0.02
Steerability · rank 11 of 46
- Unit
- IPS effect estimate, higher is better
- Range
- 0.01 to 0.03
- Sample
- 39237 observations
- Configuration
- GPT-5.6 Luna
- Measured
- 28 Sep 2026
- Not shown
- Not a success rate: the estimated effect of choosing this model as the agent's orchestrator, against an even mix of the models tested, inside Arena's own harness.
0.00
Tool grounding · rank 1 of 46
- Unit
- IPS effect estimate, higher is better
- Range
- 0.00 to 0.00
- Sample
- 3288506 observations
- Configuration
- GPT-5.6 Luna
- Measured
- 28 Sep 2026
- Not shown
- Not a success rate: the estimated effect of choosing this model as the agent's orchestrator, against an even mix of the models tested, inside Arena's own harness.
1,462
Overall · rank 14 of 44
- Unit
- Arena rating, higher is better
- Range
- 1,454 to 1,470
- Sample
- 5766 votes
- Configuration
- GPT-5.6 Luna (xhigh)
- Measured
- 13 Sep 2026
- Not shown
- Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.
1,461
Overall · rank 28 of 177
- Unit
- Arena rating, higher is better
- Range
- 1,458 to 1,465
- Sample
- 35475 votes
- Configuration
- GPT-5.6 Luna (xhigh)
- Measured
- 25 Sep 2026
- Not shown
- Not an error rate: claims that cannot be checked on the web are skipped, and preference still carries most of the weight.
1,463
Business, management and finance · rank 24 of 402
- Unit
- Arena rating, higher is better
- Range
- 1,455 to 1,471
- Sample
- 6871 votes
- Configuration
- GPT-5.6 Luna (xhigh)
- Measured
- 25 Sep 2026
- Not shown
- Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.
1,412
Creative writing · rank 69 of 407
- Unit
- Arena rating, higher is better
- Range
- 1,404 to 1,420
- Sample
- 7224 votes
- Configuration
- GPT-5.6 Luna (xhigh)
- Measured
- 25 Sep 2026
- Not shown
- Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.
1,491
Expert prompts · rank 22 of 359
- Unit
- Arena rating, higher is better
- Range
- 1,481 to 1,501
- Sample
- 4120 votes
- Configuration
- GPT-5.6 Luna (xhigh)
- Measured
- 25 Sep 2026
- Not shown
- Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.
1,446
Instruction following · rank 48 of 409
- Unit
- Arena rating, higher is better
- Range
- 1,439 to 1,452
- Sample
- 12879 votes
- Configuration
- GPT-5.6 Luna (xhigh)
- Measured
- 25 Sep 2026
- Not shown
- Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.
1,454
Overall · rank 53 of 409
- Unit
- Arena rating, higher is better
- Range
- 1,449 to 1,458
- Sample
- 35823 votes
- Configuration
- GPT-5.6 Luna (xhigh)
- Measured
- 25 Sep 2026
- Not shown
- Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.
1,435
Writing, literature and language · rank 51 of 408
- Unit
- Arena rating, higher is better
- Range
- 1,428 to 1,442
- Sample
- 9607 votes
- Configuration
- GPT-5.6 Luna (xhigh)
- Measured
- 25 Sep 2026
- Not shown
- Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.
16.0%
Requested-length error · rank 127 of 370
- Unit
- % off the requested word count, lower is better
- Configuration
- GPT-5.6 Luna (high reasoning)
- Measured
- 9 Jul 2026
- Not shown
- Not other format limits such as character counts or bullet counts.
19.0%
Requested-length error · rank 163 of 370
- Unit
- % off the requested word count, lower is better
- Configuration
- GPT-5.6 Luna (low reasoning)
- Measured
- 9 Jul 2026
- Not shown
- Not other format limits such as character counts or bullet counts.
20.0%
Requested-length error · rank 179 of 370
- Unit
- % off the requested word count, lower is better
- Configuration
- GPT-5.6 Luna (medium reasoning)
- Measured
- 9 Jul 2026
- Not shown
- Not other format limits such as character counts or bullet counts.
17.0%
Requested-length error · rank 140 of 370
- Unit
- % off the requested word count, lower is better
- Configuration
- GPT-5.6 Luna (no reasoning)
- Measured
- 9 Jul 2026
- Not shown
- Not other format limits such as character counts or bullet counts.
15.0%
Requested-length error · rank 119 of 370
- Unit
- % off the requested word count, lower is better
- Configuration
- GPT-5.6 Luna (xhigh)
- Measured
- 9 Jul 2026
- Not shown
- Not other format limits such as character counts or bullet counts.
0.37
Style adherence · rank 84 of 370
- Unit
- score from 0 to 1, higher is better
- Configuration
- GPT-5.6 Luna (high reasoning)
- Measured
- 9 Jul 2026
- Not shown
- Not brand voice on your own examples: UGI's prompts are private and lean towards creative writing.
0.37
Style adherence · rank 90 of 370
- Unit
- score from 0 to 1, higher is better
- Configuration
- GPT-5.6 Luna (low reasoning)
- Measured
- 9 Jul 2026
- Not shown
- Not brand voice on your own examples: UGI's prompts are private and lean towards creative writing.
0.38
Style adherence · rank 65 of 370
- Unit
- score from 0 to 1, higher is better
- Configuration
- GPT-5.6 Luna (medium reasoning)
- Measured
- 9 Jul 2026
- Not shown
- Not brand voice on your own examples: UGI's prompts are private and lean towards creative writing.
0.36
Style adherence · rank 136 of 370
- Unit
- score from 0 to 1, higher is better
- Configuration
- GPT-5.6 Luna (no reasoning)
- Measured
- 9 Jul 2026
- Not shown
- Not brand voice on your own examples: UGI's prompts are private and lean towards creative writing.
0.38
Style adherence · rank 58 of 370
- Unit
- score from 0 to 1, higher is better
- Configuration
- GPT-5.6 Luna (xhigh)
- Measured
- 9 Jul 2026
- Not shown
- Not brand voice on your own examples: UGI's prompts are private and lean towards creative writing.
53.7
Writing score · rank 137 of 370
- Unit
- score out of 100, higher is better
- Configuration
- GPT-5.6 Luna (high reasoning)
- Measured
- 9 Jul 2026
- Not shown
- Not business copy quality: UGI's prompts are private and lean towards creative writing, and models that often refuse get no score.
47.0
Writing score · rank 170 of 370
- Unit
- score out of 100, higher is better
- Configuration
- GPT-5.6 Luna (low reasoning)
- Measured
- 9 Jul 2026
- Not shown
- Not business copy quality: UGI's prompts are private and lean towards creative writing, and models that often refuse get no score.
49.2
Writing score · rank 160 of 370
- Unit
- score out of 100, higher is better
- Configuration
- GPT-5.6 Luna (medium reasoning)
- Measured
- 9 Jul 2026
- Not shown
- Not business copy quality: UGI's prompts are private and lean towards creative writing, and models that often refuse get no score.
41.0
Writing score · rank 209 of 370
- Unit
- score out of 100, higher is better
- Configuration
- GPT-5.6 Luna (no reasoning)
- Measured
- 9 Jul 2026
- Not shown
- Not business copy quality: UGI's prompts are private and lean towards creative writing, and models that often refuse get no score.
52.8
Writing score · rank 142 of 370
- Unit
- score out of 100, higher is better
- Configuration
- GPT-5.6 Luna (xhigh)
- Measured
- 9 Jul 2026
- Not shown
- Not business copy quality: UGI's prompts are private and lean towards creative writing, and models that often refuse get no score.
