Models / step-3

Stepfun

step-3

6 published results from 1 source. Each card shows where the number comes from and what it does not measure. The overall leaderboard combines them; here each stands alone.

Provider
Stepfun
Sources
1
Our benchmarks
0
Price per million tokens
Not listed on OpenRouter

Reported by others

1,358
Business, management and finance · rank 171 of 402
Unit
Arena rating, higher is better
Range
1,341 to 1,375
Sample
1138 votes
Configuration
step-3
Measured
25 Sep 2026
Not shown
Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.
1,310
Creative writing · rank 180 of 407
Unit
Arena rating, higher is better
Range
1,289 to 1,332
Sample
766 votes
Configuration
step-3
Measured
25 Sep 2026
Not shown
Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.
1,364
Expert prompts · rank 155 of 359
Unit
Arena rating, higher is better
Range
1,328 to 1,399
Sample
262 votes
Configuration
step-3
Measured
25 Sep 2026
Not shown
Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.
1,344
Instruction following · rank 183 of 409
Unit
Arena rating, higher is better
Range
1,330 to 1,359
Sample
1631 votes
Configuration
step-3
Measured
25 Sep 2026
Not shown
Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.
1,349
Overall · rank 199 of 409
Unit
Arena rating, higher is better
Range
1,342 to 1,357
Sample
6267 votes
Configuration
step-3
Measured
25 Sep 2026
Not shown
Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.
1,318
Writing, literature and language · rank 199 of 408
Unit
Arena rating, higher is better
Range
1,302 to 1,333
Sample
1375 votes
Configuration
step-3
Measured
25 Sep 2026
Not shown
Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.