96.4
The large value shows the leading mean for this attempt. Rankings are calculated across repeated attempts rather than from this result alone.
- Attempt
- adv-3003
- Model
- anthropic/claude-sonnet-4.6
- Questions asked
- 110
- Seed
- 3003
- Fingerprint
- ecec4d8ad71b98a91eaf4490
- When
- 2026-09-10T07:15:50.754395+00:00
- Duration
- 12.6 min
- Cost
- 3.79 USD billed
- Tokens
- 88.2k prompt · 570.5k completion · 658.7k total
openai/gpt-5.4
mean on this sampled test
· 0.249 USD billed
· 14.6k / 14.2k / 28.8k tok
· 2.0 s / question
91.4
openai/o4-mini
mean on this sampled test
· 0.285 USD billed
· 14.6k / 61.0k / 75.6k tok
· 4.5 s / question
92.9
qwen/qwen3-max
mean on this sampled test
· 0.140 USD billed
· 15.2k / 32.9k / 48.1k tok
· 5.1 s / question
87.9
google/gemini-2.5-pro
mean on this sampled test
· 2.36 USD billed
· 14.1k / 235.5k / 249.6k tok
· 19.8 s / question
71.4
deepseek/deepseek-r1-0528
mean on this sampled test
· 0.473 USD billed
· 14.0k / 211.0k / 225.0k tok
· 55.2 s / question
73.6
anthropic/claude-sonnet-4.6
mean on this sampled test
· 0.285 USD billed
· 15.7k / 15.9k / 31.6k tok
· 3.0 s / question
96.4