OpenAI AI Updates: October 31, 2026

1. OpenAI Adds GPT-6 Rosalind, a Life Sciences Reasoning Model, to the API Under Trusted Access

OpenAI added GPT-6 Rosalind (gpt-6-rosalind) to its API changelog on October 30. The entry describes the model as offering specialized biological reasoning with scientific tool use for complex life sciences research. Access is limited to organizations approved for Rosalind Research through OpenAI’s trusted-access program, and is restricted to internal research by authorized users. OpenAI’s pricing page lists gpt-6-rosalind at $15 per 1M input tokens and $75 per 1M output tokens for inputs up to 272K tokens, rising to $30 and $112.50 above that, with cached input at $1.50 and $3.00 respectively. The earlier gpt-rosalind-research and gpt-rosalind-discovery models remain listed at $5 input and $25 output, and OpenAI says eligible organizations will receive new GPT-Rosalind models as they are released. Source

2. OpenAI Will Honor a Shared robots.txt Signal for AI Training Preferences

OpenAI said it will respect a shared robots.txt signal that lets website operators state their AI training preferences to participating providers. OpenAI’s existing crawler documentation still lists disallowing GPTBot as the way to keep a site’s content out of foundation model training, with OAI-SearchBot, ChatGPT-User, and OAI-AdsBot governed separately, and it does not yet name the new signal. The full post could not be retrieved, so the name of the signal, the other participating providers, and the date it takes effect were not confirmed. Source

3. OpenAI Publishes a Community Safety Report and Expands Violence Prevention Safeguards in ChatGPT

OpenAI published a community safety report together with new protections for teens. According to the summary, OpenAI is strengthening violence prevention in ChatGPT with broader safeguards and earlier interventions, and is committing to publish public safety reports going forward. The update follows the August launch of ChatGPT for Teens and the teen usage report OpenAI released in early October. The full post could not be retrieved, so the figures in the report and the mechanics of the new interventions were not confirmed. Source

4. OpenAI Compiles Court Brief Excerpts Arguing Fair Use Covers AI Training

OpenAI published a post titled “Fair use benefits everyone” that collects excerpts from court briefs filed by creators, academics, accessibility advocates, and legal experts arguing that fair use supports AI training and innovation. The post is a policy piece and does not announce product or model changes. The full post could not be retrieved, so the specific cases in which the briefs were filed were not confirmed. Source

5. Credit Saison Reports 954% ROI From a ChatGPT Enterprise Pilot

OpenAI published a customer story on Credit Saison, the Japanese financial services company, whose ChatGPT Enterprise pilot delivered a 954% return on investment according to the summary. The post describes how company leaders and employees turned the pilot results into company-wide adoption. The full post could not be retrieved, so the pilot size, timeframe, and use cases were not confirmed. Source