Skip to content

Benchmark

Artificial Analysis τ-Voice · Agentic Performance

Scores are percentages. Higher is better.

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

Score distribution

56.8%
Mean
68.6%
Median
18.9%
Std dev
30.1%
10th percentile
71.6%
90th percentile
30.1%
Min
30.1%71.6%
The band covers the middle 80% of scores. The line marks the median.

Leaderboard

#ModelScore
1Grok Voice 3.0October 13 release article · Artificial Analysis τ-Voice customer-support tasks· Reported by SpaceXAI, Artificial Analysis· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.71.6%
2Gemini 3.8 Live Extended Thinkinghigh thinking · Gemini API · pass@1· Reported by Google DeepMind, Google· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.68.6%
3Gemini 3.8 LiveGemini API · pass@1 · thinking level not printed on chart· Reported by Google DeepMind· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.30.1%

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.