OpenAI said earlier versions, including GPT-5.2 Instant, would sometimes refuse questions that could have been answered safely or respond with what users described as overly cautious or preachy language.
With GPT-5.3 Instant, OpenAI said the model “significantly reduces unnecessary refusals” and tones down overly defensive or moralising preambles before answering.
OpenAI said the model is also available to developers through the API under the name “gpt-5.3-chat-latest”.
GPT-5.3 Instant reduces hallucination rates across multiple domains, including higher-stakes areas such as medicine, law and finance.
Ключевые утверждения в источнике B
For OpenAI this makes sense, because it will also save it money on inference costs,” the analyst said.
Of course, not everyone will be happy with the changes, but it’s impossible to please them all, and OpenAI is facing massive pressure from Google.” The company said GPT-5.3 Instant will replace GPT-5.2 Instant as the de…
GPT‑5.3 Instant, on the other hand, gets right into the response.” The company said its tests show that GPT-5.3 Instant shows improvements in factual reliability as well as changes in the tone of its responses.
It said hallucinations have also declined by around 27% when the model uses the web to inform its responses, and by 20% when relying only on the knowledge it was trained on.
Текстовые доказательства
Доказательства из источника A
ключевое утверждение
With GPT-5.3 Instant, OpenAI said the model “significantly reduces unnecessary refusals” and tones down overly defensive or moralising preambles before answering.
Ключевое утверждение, на котором строится подача материала.
ключевое утверждение
OpenAI said the model is also available to developers through the API under the name “gpt-5.3-chat-latest”.
Ключевое утверждение, на котором строится подача материала.
выборочный акцент
In internal evaluations focused on higher-stakes topics, OpenAI said GPT-5.3 Instant reduced hallucination rates by 26.8 per cent when using web access and by 19.7 per cent when relying onl…
Возможное выборочное акцентирование отдельных аспектов истории.
Доказательства из источника B
ключевое утверждение
For OpenAI this makes sense, because it will also save it money on inference costs,” the analyst said.
Ключевое утверждение, на котором строится подача материала.
ключевое утверждение
GPT‑5.3 Instant, on the other hand, gets right into the response.” The company said its tests show that GPT-5.3 Instant shows improvements in factual reliability as well as changes in the t…
Ключевое утверждение, на котором строится подача материала.
выборочный акцент
Of course, not everyone will be happy with the changes, but it’s impossible to please them all, and OpenAI is facing massive pressure from Google.” The company said GPT-5.3 Instant will rep…
Возможное выборочное акцентирование отдельных аспектов истории.
Доказательства паттернов искажения
Источник A · Эффект фрейминга
In internal evaluations focused on higher-stakes topics, OpenAI said GPT-5.3 Instant reduced hallucination rates by 26.8 per cent when using web access and by 19.7 per cent when relying onl…
Of course, not everyone will be happy with the changes, but it’s impossible to please them all, and OpenAI is facing massive pressure from Google.” The company said GPT-5.3 Instant will rep…