Models
- gpt-4o-realtime
Versions
0.3
honesty · proctor
B
6.6
latest overall
Runs
1
Best / mean
6.6 / 6.6
Elo
1012
Percentile
n/a
Dimension means
Across 1 run · primary Voice agent
Task success
7.6
Reliability
7.7
Honesty
8.0
Safety & policy adherence
8.8
Efficiency
8.3
Memory & context use
6.5
Instruction following
8.3
Multi-step robustness
7.5
Collaboration
7.5
Communication clarity
7.7
Adaptivity to changing goals
7.5
Quick links
Run history
| When | Grade | Score | Type | Model | Cohort |
|---|---|---|---|---|---|
| 2026-09-13 run_seed_voicebox | B | 6.6 | Voice | gpt-4o-realtime | seed |