OpenAI says it added computer-use agents, stronger clinical search and more than 20 healthcare plugins to ChatGPT for Healthcare. No prediction venue prices that rollout, so there are no current odds, no volume and nothing moved. Agents still need workspace limits and human review before acting in live systems. Not financial advice. Markets move fast.
Update, Nov 16, 15:31 UTC · via OpenAI News
OpenAI News reported that GPT-6 Astra was evaluated on clinician EHR inbox tasks shaped by physician input, with 99% of 325 reviewed tasks rated safe and physicians selecting "very helpful" or "extremely helpful" in 93% of 646 assessments. On HealthAdminBench, GPT-6 Astra earned 94.3% of available rubric points while GPT-6 Luna earned 91.8% despite costing 98% less at 8.5 cents per task.
OpenAI News reported that a CVS Health production workflow cleared a months-long document-sending backlog over a weekend and another handles system steps for tens of thousands of consults each day, with CVS Health open sourcing its Agent Kit built with OpenAI's Agents SDK. Lagan Khare, Executive Director, Engineering, CVS Health, said "Less time on administrative work means more time for care. That’s a win for our teams and everyone we serve.".
The pitch is less chat, more doing.
OpenAI said computer-use agents can read screens, click, type and move between applications to carry work across software teams already use. The company named Abridge, Ambience, Cigna Healthcare and CVS Health as early deployers using agents to process referrals, move documents and assemble care information. It said customers can limit which applications agents touch and decide which steps need human review.
For clinical answers, OpenAI said it upgraded how ChatGPT finds, evaluates and cites medical evidence in ChatGPT for Healthcare and ChatGPT for Clinicians. OpenAI said physicians in a blind test of real-world questions across specialties filed more than 1,300 reviews and rated the new system higher on all eight measures, including factual correctness, safety, actionability and source authoritativeness.
The plugin push is where the context gets crowded. OpenAI said the new healthcare plugins are live now, including ChartGPT access inside the EHR workflow, with names like Aiva, Amion, Availity, Elation Health, IMO Health, Metriport, Oracle HDI, Redox, Stedi, TriZetto, Turquoise and UpToDate Expert AI. Luma Health said its ChatGPT plugin alone exposes 59 actions for prep, rescheduling, waitlists, outreach and balance follow-up, per its announcement. An earlier public healthcare search plugin brought together nine apps for biomedical research, trials, medication information, Medicare data and provider records, per OpenAI Help Center.
That contrast matters. Luma talks about running full clinic workflows from one prompt, while OpenAI talks about access controls, audit logs and required human review for clinical and administrative work. OpenAI did not name the agent models, share error rates or price the healthcare workspace in this announcement. It said content shared with ChatGPT for Healthcare is not used to train models and said HIPAA workflows need eligible services under a BAA plus required security and retention settings.
For access, OpenAI said computer-use agents in ChatGPT Work and Codex plus improved search are available now, subject to workspace settings, while workspace admins must enable the new plugins and individual clinicians can use select plugins through free ChatGPT for Clinicians accounts for eligible verified U.S. clinicians. So who owns a mistake when an agent clicks through three systems and drafts the wrong MRI order?





