Health data and physical agents arrive in the same week
ChatGPT Health and OpenAI for Healthcare, and The Deploy Log for this week. Every price and claim read off the publisher's own page, with a call on each.
OpenAI
ChatGPT Health
DEPLOYif you use ChatGPT for health questions
WHAT SHIPPEDA dedicated health space in ChatGPT with separate storage, memories, and encryption, launching to U.S. users 18 and older on web and iOS. Over 230 million people already ask health questions on ChatGPT weekly.
THE CATCHHealth conversations are not used to train foundation models, and health memories never flow back into non-Health chats. Medical record integrations and some apps are U.S. only, and Apple Health requires iOS.
IF YOU RUNa consumer app where users already discuss symptoms, medications, or wellness data
SOURCEOpenAI, Introducing ChatGPT Health
OpenAI
OpenAI for Healthcare
DEPLOYif you run a HIPAA-regulated healthcare organization
WHAT SHIPPEDChatGPT for Healthcare, available starting today and rolling out to AdventHealth, Baylor Scott & White Health, Boston Children's Hospital, Cedars-Sinai Medical Center, HCA Healthcare, Memorial Sloan Kettering Cancer Center, Stanford Medicine Children's Health, and UCSF.
THE CATCHGPT-5.2 models outperform human baselines across every role measured in GDPval, and content shared with ChatGPT for Healthcare is not used to train models. A BAA with OpenAI supports HIPAA-compliant use.
IF YOU RUNa hospital or clinic where clinicians draft discharge summaries, referral letters, or patient instructions
SOURCEOpenAI, Introducing OpenAI for Healthcare
The Deploy Log
Models and APIs
| DEPLOYif you build robots or video analytics agents NVIDIA Cosmos Reason 2. NVIDIA released Cosmos Reason 2, an open reasoning vision language model for physical AI that tops the Physical AI Bench and Physical Reasoning leaderboards as the #1 open model for visual understanding. Hugging Face |
| DEPLOYif you serve Arabic language users Falcon-H1-Arabic. Falcon-H1-Arabic 3B, 7B, and 34B models outperform all SOTA models of similar sizes and sometimes bigger, with the 34B reaching approximately 75% on OALL and surpassing Llama-3.3-70B. Hugging Face |
| DEPLOYif you run large codebases with incident risk Datadog Codex review. Datadog deployed Codex across its engineering workforce, with more than 1,000 engineers using it regularly, and Codex found more than 10 cases, roughly 22% of incidents examined, where its feedback would have made a difference. OpenAI |
| DEPLOYif you run enterprise customer service agents Netomi agentic systems. Netomi pairs GPT-4.1 for low-latency tool use with GPT-5.2 for multi-step planning, sustaining sub-three-second responses with 98% intent classification accuracy during traffic spikes above 40,000 concurrent customer requests per second. OpenAI |
| WATCHif you build voice agents Tolan voice AI. Tolan built a voice-first AI companion with GPT-5.1, cutting speech initiation time by over 0.7 seconds, dropping memory recall misses by 30%, and raising next-day user retention more than 20%. OpenAI |
Robotics, hardware and chips
| WATCHif you prototype physical AI agents DGX Spark and Reachy Mini. NVIDIA showed how to build a personal office robot agent using DGX Spark with Reachy Mini, combining NVIDIA Nemotron 3 Nano for reasoning, Nemotron Nano 2 VL for vision, and ElevenLabs for text-to-speech. Hugging Face |
Left out
- Bookmark domains: it did not carry a number that moved
- One build per branch: it did not carry a number that moved
- bash-tool: it did not carry a number that moved
- Claude Code via AI Gateway: it did not carry a number that moved
- Vercel Agent code guidelines: it did not carry a number that moved
- Secure Compute self-serve: it did not carry a number that moved
- Stargate SB Energy partnership: the announcement hedges, so nothing has deployed yet
The pattern
OpenAI shipped ChatGPT Health and ChatGPT for Healthcare in the same week that NVIDIA released Cosmos Reason 2 and demonstrated a DGX Spark with Reachy Mini office robot agent. The health products separate storage, memories, and encryption for 230 million weekly ChatGPT users and for hospitals including AdventHealth, Baylor Scott & White Health, Boston Children's Hospital, Cedars-Sinai Medical Center, HCA Healthcare, Memorial Sloan Kettering Cancer Center, Stanford Medicine Children's Health, and UCSF, while GPT-5.2 models outperform human baselines in GDPval and a BAA supports HIPAA-compliant use. The physical AI releases make Cosmos Reason 2 the #1 open model on the Physical AI Bench and Physical Reasoning leaderboards and pair NVIDIA Nemotron 3 Nano, Nemotron Nano 2 VL, and ElevenLabs in a personal office robot agent.
The same week also shows production systems moving past model quality into latency, retention, and incident impact. Datadog's Codex review found more than 10 cases, roughly 22% of incidents examined, where feedback would have made a difference across more than 1,000 engineers. Netomi pairs GPT-4.1 with GPT-5.2 to hold sub-three-second responses and 98% intent classification accuracy above 40,000 concurrent customer requests per second. Tolan uses GPT-5.1 to cut speech initiation time by over 0.7 seconds, drop memory recall misses by 30%, and raise next-day user retention more than 20%. Falcon-H1-Arabic 3B, 7B, and 34B reach approximately 75% on OALL and surpass Llama-3.3-70B, showing that smaller open models now compete with larger ones while enterprise deployments measure seconds and retention points.