Skip to content

Benchmark

Full Duplex Bench v3 · Response Quality

Scores are percentages. Higher is better.

Models and systems measured
3
Scores recorded
3
Sources
1
IQ weight
No weight published

Score distribution

86.3%
Mean
88%
Median
3.9%
Std dev
81%
10th percentile
90%
90th percentile
81%
Min
81%90%
The band covers the middle 80% of scores. The line marks the median.

Leaderboard

#Model or systemScore
1GPT-Live 1OpenAI release chart · spoken tool-using requests · answer/reference-intent match · GPT-Live-1 backend: GPT-5.6 Terra · low reasoning· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.90%
2GPT Realtime 2OpenAI release chart · spoken tool-using requests · answer/reference-intent match· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.81%

Not ranked

The source marks these results as descriptive, not capability scores, so we list them by name instead of ranking them.

Model or systemScore
GPT-Realtime-2.1OpenAI release chart · spoken tool-using requests · answer/reference-intent match· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.88%

Data credits: BenchBoard by SYuan03 (github.com/SYuan03/benchboard). MIT covers its code and data organization; no results licence is inferred. Numeric relays retain the original stating-source links.