Chaptered summary is still being generated for this document. Showing a heuristic brief in the meantime.
Summary
1,550-word document condensed to 104 words. OpenAI · Aug 3, 2026
TL;DR
“GPT-Live-1 and GPT-Live-1 mini are a new generation of voice models de- signed to make conversations with AI feel more natural and intelligent. These models — enabled by research advances — are full-duplex, meaning they can listen and respond continuously instead of waiting for a clearly defined turn to end.”
Top benchmarks
| Benchmark | Variant | Score |
|---|---|---|
| voice_native_synthetic_prompts | hate, accuracy | 100.0% |
| voice_native_synthetic_prompts | self_harm, accuracy | 98.0% |
| voice_native_synthetic_prompts | hate, accuracy | 98.0% |
| voice_native_synthetic_prompts | illicit_behavior, accuracy | 97.0% |
| voice_native_synthetic_prompts | sexual, accuracy | 97.0% |
| voice_native_synthetic_prompts | illicit_behavior, accuracy | 97.0% |
| voice_native_production_prompts | sexual, accuracy | 97.0% |
| voice_native_production_prompts | illicit_behavior, accuracy | 97.0% |
Showing top 8 of 62. See full list below.
Capability claim
- “We trained these new models to respond safely , using the same infrastruc- ture we rely on for training our flagship models.”
Every italicized passage is a verbatim substring of the source document (checked deterministically after extraction). Field selection is heuristic — some quotes may lack surrounding context and some claims may be absent if no matching pattern appeared. For citation, open the source: original model card · source SHA 9a2e9dd82dcb · version dated Aug 3, 2026.
Extracted Evaluations(62 results)
Sort by:0/62 rows fully reproducible (0%)
| Benchmark | Category | State | Score | Setup | Source |
|---|---|---|---|---|---|
voice_native_synthetic_prompts/ hate | other | scored | 1.0 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_synthetic_prompts/ self_harm | other | scored | 1.0 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_synthetic_prompts/ hate | other | scored | 1.0 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_synthetic_prompts/ illicit_behavior | other | scored | 1.0 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_synthetic_prompts/ sexual | other | scored | 1.0 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_synthetic_prompts/ illicit_behavior | other | scored | 1.0 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_production_prompts/ sexual | other | scored | 1.0 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_production_prompts/ illicit_behavior | other | scored | 1.0 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_production_prompts/ sexual | other | scored | 1.0 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_synthetic_prompts/ gore | other | scored | 1.0 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_synthetic_prompts/ personal_data | other | scored | 1.0 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_synthetic_prompts/ personal_data | other | scored | 1.0 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_production_prompts/ personal_data | other | scored | 1.0 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_synthetic_prompts/ self_harm | other | scored | 1.0 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_production_prompts/ sexual | other | scored | 1.0 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_synthetic_prompts/ gore | other | scored | 1.0 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_production_prompts/ self_harm | other | scored | 1.0 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_production_prompts/ personal_data | other | scored | 1.0 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_production_prompts/ personal_data | other | scored | 0.9 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_production_prompts/ sexual | other | scored | 0.9 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_production_prompts/ personal_data | other | scored | 0.9 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_synthetic_prompts/ sexual | other | scored | 0.9 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_production_prompts/ illicit_behavior | other | scored | 0.9 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_production_prompts/ self_harm | other | scored | 0.9 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_synthetic_prompts/ emotional_reliance | other | scored | 0.9 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_production_prompts/ mental_health | other | scored | 0.9 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_production_prompts/ mental_health | other | scored | 0.9 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_synthetic_prompts/ personal_data | other | scored | 0.9 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_production_prompts/ self_harm | other | scored | 0.9 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_synthetic_prompts/ emotional_reliance | other | scored | 0.9 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_production_prompts/ emotional_reliance | other | scored | 0.9 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_synthetic_prompts/ hate | other | scored | 0.9 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_synthetic_prompts/ sexual | other | scored | 0.8 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_production_prompts/ mental_health | other | scored | 0.8 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_synthetic_prompts/ mental_health | other | scored | 0.8 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_production_prompts/ emotional_reliance | other | scored | 0.8 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_synthetic_prompts/ hate | other | scored | 0.8 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_production_prompts/ self_harm | other | scored | 0.8 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_synthetic_prompts/ mental_health | other | scored | 0.8 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_synthetic_prompts/ gore | other | scored | 0.8 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_synthetic_prompts/ personal_data | other | scored | 0.8 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_production_prompts/ mental_health | other | scored | 0.8 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_production_prompts/ emotional_reliance | other | scored | 0.8 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_production_prompts/ emotional_reliance | other | scored | 0.8 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_synthetic_prompts/ sexual | other | scored | 0.8 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_production_prompts/ illicit_behavior | other | scored | 0.7 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_synthetic_prompts/ self_harm | other | scored | 0.7 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_synthetic_prompts/ emotional_reliance | other | scored | 0.7 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_synthetic_prompts/ emotional_reliance | other | scored | 0.7 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_synthetic_prompts/ self_harm | other | scored | 0.7 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_synthetic_prompts/ illicit_behavior | other | scored | 0.6 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_synthetic_prompts/ illicit_behavior | other | scored | 0.6 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_synthetic_prompts/ gore | other | scored | 0.6 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_production_prompts/ illicit_behavior | other | scored | 0.6 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_synthetic_prompts/ mental_health | other | scored | 0.6 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
voice_native_synthetic_prompts/ mental_health | other | scored | 0.5 accuracy | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
Preparedness Framework/ biological_chemical_risk | other | mentioned | — | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
Preparedness Framework/ biological_chemical_risk | other | mentioned | — | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
Preparedness Framework/ ai_self_improvement | other | mentioned | — | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
Preparedness Framework/ cybersecurity | other | mentioned | — | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
Preparedness Framework/ cybersecurity | other | mentioned | — | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |
Preparedness Framework/ ai_self_improvement | other | mentioned | — | missing: shot countmissing: methodmissing: languagemissing: training state | self-reported |