OpenAI

GPT-Live
System Card
2026-07-08
Change log
        • August 4, 2026: We identified a configuration mismatch in the safety
          evaluation setup: the results had been generated using a backend con-
          figuration that did not match the final model release. We re-ran the full
          evaluation using the correct configuration and updated the reported
          results. Some values remained unchanged after rerunning, while others
          were corrected. We’ve added an appendix to show the original and cor-
          rected values side by side. The overall safety conclusions are unchanged;
          none of the regressions led to numbers below our safety launch stan-
          dards.

1 Introduction
      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. This means they can follow pauses, interruptions, and
      changes in pace, and decide in the moment whether to respond or keep
      listening.
      GPT-Live-1 will be the default voice model for paid users, while GPT-Live-1
      mini will be the default model for free users.
      The most important things to know about our safety work for this launch
      are that:

    1. We trained these new models to respond safely, using the same infrastruc-
       ture we rely on for training our flagship models. They can also delegate
       more complex work to our other models, and when they do, the resulting
       work will reflect the safety training of the underlying model that is doing
       that work. The evaluations in this card describe how the new GPT-Live-1
       models perform with delegation, matching the deployment context.

    2. These models have system-level safety integrations that are designed to
       be on par with the existing safety stack for text models, while adopting
       some new safeguards specifically for the new voice modality: inputs and
       generated outputs are checked as the conversation unfolds; when potentially
       unsafe content is detected, the system can steer or interrupt the response,
       play a spoken safety message, provide support resources in text, or, in
       higher-risk cases, end the voice conversation.

1                                                                                     OpenAI
        3. As part of our safety work for this launch, we built new evaluations that
           focus specifically on the distinctive ways that people use voice models, as
           opposed to text-based chats, and on observations from real-world use of the
           existing Advanced Voice Mode. We report those results below.

        4. The same monitoring, review, and enforcement infrastructure that we use
           for our text models also applies to the GPT-Live models, enabling us to
           measure prevalence, detect abuse, and enforce our safety policies.

2 Model Data and Training
           Like OpenAI’s other models, GPT-Live-1 and GPT-Live-1 mini were trained
           on diverse datasets, including information that is publicly available on the
           internet, information that we partner with third parties to access, and in-
           formation that our users or human trainers and researchers provide or gen-
           erate. Our data processing pipeline includes rigorous filtering to maintain
           data quality and mitigate potential risks. We use advanced data filtering pro-
           cesses to reduce personal information from training data. We also employ
           safety classifiers to help prevent or reduce the use of harmful or sensitive
           content, including explicit materials such as sexual content involving a
           minor.
           Note that comparison values from previously launched models are from the
           latest versions of those models, so may vary slightly from values published
           at launch for those models.1

3 Model Safety

1
    GPT-Live-1 and GPT-Live-1 mini are intended to be used in accordance with OpenAI’s Usage Policies, Service
    Terms, and Terms of Use. These policies apply universally to OpenAI services and are designed to ensure safe
    and responsible usage of AI technology. You can review OpenAI’s Usage Policies at openai.com/policies/usage-
    policies/ .

    If you need assistance with respect to GPT-Live-1 and GPT-Live-1 mini, you can find further information on
    OpenAI’s website (openai.com), or you can contact OpenAI Support by opening the chat bubble icon displayed
    at the bottom-right of help.openai.com.
    A list of the languages that ChatGPT currently supports can be found here.

2                                                                                                        OpenAI
3.1 Voice-Native Evaluations for Disallowed Content
             August 4, 2026: We identified a configuration mismatch in the evaluation
             setup. We re-ran the full evaluation using the correct configuration, and
             the results reported in this section have been updated accordingly. We’ve
             added an appendix to show the original and corrected values side by side.

3.1.1 Voice-Native Evaluations: Production Prompts

             In these newly developed evaluations, we use real audio examples from
             users who have chosen to share their voice interactions to help improve
             our models. Before these examples are used, they are processed through
             our privacy and eligibility safeguards, including checks for user permis-
             sions and deletion/opt-out status, filtering of ineligible data, and steps to
             reduce personal information through PII scrubbing and de-identification.
             We then transcribe the audio, generate the model’s response, and evaluate
             that response for safety.
             We compare the new GPT-Live models to their respective predecessors, the
             models that power Advanced Voice Mode (AVM).
             These evaluations are not prevalence weighted, meaning they do not reflect
             rates of safety performance we see in real usage. Instead, these evaluations
             were meant to be difficult. For each category below, the evaluation is built
             around cases in which the existing AVM models were not yet giving ideal
             responses.

Table 1: Voice-Native Evaluations: Production Prompts

                                   AVM            GPT-Live-1        AVM mini       GPT-Live-1 mini

    Sexual                         0.96              0.97             0.97              0.92
    Illicit behavior               0.74              0.95             0.60              0.91
    Mental health                  0.90              0.90             0.78              0.85
    Personal data                  0.96              0.96             0.95              0.96
    Emotional reliance             0.88              0.82             0.78              0.78
    Self-harm                      0.89              0.97             0.81              0.95

             We observe that the GPT-Live models generally provide equal or better
             safety performance across these adversarially selected prompts than AVM
             models. GPT-Live-1 shows a slight regression on emotional reliance from
             0.88 to 0.82, and GPT-Live-1 mini shows a slight regression on sexual content
             from 0.97 to 0.92.

3                                                                                              OpenAI
3.1.2 Voice-Native Evaluations: Synthetic Prompts

             The below evaluations are similar to the ones above, except that for these
             we synthetically generate audio prompts to target challenging edge cases.
             The text for these prompts is generated from safety policies and related
             guidance, targeting specific safety categories, policy boundaries, and dif-
             ficult cases. This approach allows us to deliberately cover a broad range
             of scenarios, including rare or hard-to-sample safety-relevant situations.
             We then convert the prompts into speech and use them as audio inputs,
             allowing us to assess whether models apply the intended safety behavior
             when safety-relevant content is presented in spoken form. These evalua-
             tions focus more intensively on the areas where we believe our existing
             models are least likely to give ideal responses. As a result, the numbers in
             the table below are likewise not a guide to safety performance across all
             production traffic.

Table 2: Voice-Native Evaluations: Synthetic Prompts

                                   AVM            GPT-Live-1        AVM mini       GPT-Live-1 mini

    Sexual                         0.76                0.95           0.85              0.95
    Illicit behavior               0.63                0.98           0.63              0.95
    Mental health                  0.57                0.85           0.47              0.80
    Personal data                  0.90                0.98           0.79              0.98
    Emotional reliance             0.72                0.93           0.72              0.87
    Self-harm                      0.72                0.96           0.71              0.95
    Hate                           0.87                1.00           0.82              0.99
    Gore                           0.61                0.93           0.80              0.90

             We observe that the GPT-Live models uniformly provide equal or better
             safety performance across these synthetic, adversarially selected prompts
             than AVM models.
             The synthetic evaluation is designed to assess different risk surfaces than
             the production evaluation, so performance may vary between them. The
             synthetic set tests targeted, policy-grounded adversarial scenarios, includ-
             ing rare cases that may be underrepresented in production data. Strong
             performance on this set indicates that safety training is transferring for
             clear, intentionally constructed risks, but may not translate to production
             behavior. The production set reflects real-world user behavior and often in-
             cludes more ambiguous or borderline context, longer interaction histories,
             and persistent attempts to steer the model toward unsafe outputs. Failures
             in the production set can be subtler and less severe, but still important. We
             therefore view the two benchmarks as complementary.

4                                                                                              OpenAI
4 Red Teaming
    OpenAI worked with a team of internal and external red teamers across
    languages to stress test the models’ safety training with no system level mit-
    igations. We started very early in the process to baseline performance with
    no model safety training, and ultimately completed two additional rounds
    of testing as we continued to improve safeguards. Red teaming spanned
    multiple categories including child-coded voice, impersonation, speaker
    identification, sensitive train identification, self-harm, emotional reliance,
    scams and manipulation, and audio-specific perturbations. Early findings
    were used to prioritize risk areas to focus mitigation efforts on (e.g. sexual
    content, emotional reliance, and self harm) while validating some other
    areas that were policy-compliant by default (e.g. voice cloning and imper-
    sonations). Consequent follow-up rounds validated that the mitigations we
    built to address the identified issues are robust.

5 Preparedness Framework
    The Safety Advisory Group reviewed this launch and determined that nei-
    ther GPT-Live-1 nor GPT-Live-1 mini, when operating without delegation,
    could plausibly be considered High in any of our Preparedness Framework’s
    Tracked Categories – Biological and Chemical Risk, AI Self-Improvement,
    or Cybersecurity.

     • For the Biological and Chemical domain, where GPT-Live models can
       use our highly capable flagship models, the GPT-Live experience inher-
       its the safeguards of those underlying models. In addition, we’ve built
       automated monitors that may interrupt and end the call when poten-
       tially harmful conversations are detected, or degrade user experience
       for repeated abuse. We also take actor level enforcement actions where
       necessary.

     • For the Cybersecurity domain, delegated work will likewise receive the
       safeguards associated with the model to which work is delegated. In
       addition, cybersecurity risk from the GPT-Live models themselves is
       highly constrained at launch because these models lack broad access
       to tools independently of the models to which they delegate, and do
       not have code execution capability. We will reassess the cybersecurity
       safeguards posture before enabling additional tools.

     • AI Self-Improvement capability evals were not run, as GPT-Live-1 and
       GPT-Live-1 mini are less capable than GPT-5.5 Thinking across several
       intelligence evaluations.

5                                                                                    OpenAI
Appendix A: Original and corrected values for model
safety evaluations
               This section was added on August 4, 2026.
               Compared with the previously reported results, the corrected production
               evaluation scores show statistically significant regressions on illicit behavior
               for GPT-Live-1 (0.97 to 0.95) and GPT-Live-1 mini (0.94 to 0.91). On the
               synthetic evaluations, GPT-Live-1 shows statistically significant regressions
               on sexual content (0.97 to 0.95) and gore (0.97 to 0.93), while GPT-Live-1 mini
               shows statistically significant regressions on illicit behavior (0.97 to 0.95)
               and gore (0.96 to 0.90). None of the regressions fall below our safety launch
               standards.

Table 3: Voice-Native Evaluations: Production Prompts — Original and Corrected Results

    Category                AVM          GPT-Live-1     GPT-Live-1     AVM mini       GPT-Live-1 mini GPT-Live-1 mini
                                     Previously reported Corrected                  Previously reported Corrected

    Sexual                  0.96          0.97              0.97          0.97           0.95              0.92
    Illicit behavior        0.74          0.97              0.95          0.60           0.94              0.91
    Mental health           0.90          0.90              0.90          0.78           0.84              0.85
    Personal data           0.96          0.96              0.96          0.95           0.95              0.96
    Emotional reliance      0.88          0.82              0.82          0.78           0.78              0.78
    Self-harm               0.89          0.96              0.97          0.81           0.92              0.95

Table 4: Voice-Native Evaluations: Synthetic Prompts — Original and Corrected Results

 Category                 AVM          GPT-Live-1     GPT-Live-1     AVM mini       GPT-Live-1 mini   GPT-Live-1 mini
                                   Previously reported Corrected                  Previously reported   Corrected

 Sexual                   0.76         0.97             0.95           0.85             0.95               0.95
 Illicit behavior         0.63         0.97             0.98           0.63             0.97               0.95
 Mental health            0.57         0.84             0.85           0.47             0.81               0.80
 Personal data            0.90         0.97             0.98           0.79             0.97               0.98
 Emotional reliance       0.72         0.91             0.93           0.72             0.89               0.87
 Self-harm                0.72         0.98             0.96           0.71             0.96               0.95
 Hate                     0.87         1.00             1.00           0.82             0.98               0.99
 Gore                     0.61         0.97             0.93           0.80             0.96               0.90

6                                                                                                                 OpenAI