Get the weekly research letter
Companies change their terms quietly. We read every version and catch what actually changed. One email a week on the changes that matter and what they mean. No account.
The document states that API access provides developers with direct access to gpt-5-thinking, gpt-5-thinking-mini, and gpt-5-thinking-nano, with the nano variant described as specifically designed for developer use cases.
This analysis describes what OpenAI's agreement states, permits, or reserves. It does not constitute a legal determination about enforceability. Regulatory applicability and practical outcomes may vary by jurisdiction, enforcement context, and individual circumstances. Read our methodology
This provision defines the model variants available to API developers, which determines the capability ceiling for applications built on the GPT-5 platform and establishes that the High capability bio/chem classification and safe-completions restrictions apply to the thinking model variants accessible via the API.
Interpretive note: The document does not detail what specific API usage restrictions, if any, are imposed by the High capability bio/chem classification on developer access to gpt-5-thinking variants, creating uncertainty about operational constraints for API users.
This provision establishes the specific model variants available to developers through the API, which governs the capabilities and associated safeguards, including safe-completions and the High bio/chem capability classification, that apply to applications built on gpt-5-thinking variants.
Cross-platform context
See how other platforms handle Developer API Access to Thinking Model Variants and similar clauses.
Compare across platforms →Monitoring
OpenAI has changed this document before.
Receive same-day alerts, structured change summaries, and monitoring for up to 25 platforms.
"In the API, we provide direct access to the thinking model, its mini version, and an even smaller and faster nano version of the thinking model, made for developers (gpt-5-thinking-nano).Excerpt from OpenAI's GPT-5 System Card
1) REGULATORY LANDSCAPE: API access to gpt-5-thinking, which carries the High capability bio/chem classification, may engage obligations under the EU AI Act for deployers who build products using this model, as deployer obligations under that regulation may apply regardless of whether the provider has implemented safeguards. In the United States, FTC guidance on AI accountability may be relevant to how developers represent the capabilities and safety properties of applications built on gpt-5-thinking. 2) GOVERNANCE EXPOSURE: Medium. Developers accessing gpt-5-thinking via the API inherit the model's High capability bio/chem classification and associated safeguards, but the document does not specify what additional obligations or restrictions apply to API users building applications in bio/chem-adjacent domains. This creates a governance gap that developer legal teams should investigate via OpenAI's API usage policies. 3) JURISDICTION FLAGS: EU/EEA developers face heightened obligations as deployers of a general-purpose AI model with disclosed systemic risk characteristics under the EU AI Act. Developers in the United States operating in dual-use research or life sciences sectors should assess whether API deployment of gpt-5-thinking triggers any sector-specific regulatory obligations. 4) CONTRACT AND VENDOR IMPLICATIONS: API developers should review OpenAI's API terms of service to confirm whether restrictions associated with the High capability bio/chem classification are contractually specified and whether liability for misuse is allocated between OpenAI and the API customer. Standard developer agreements may not explicitly address the implications of deploying a High capability classified model. 5) COMPLIANCE CONSIDERATIONS: Developers building applications using gpt-5-thinking variants should document the model's risk classification in their own AI governance records and assess whether their application's use case is consistent with the safeguards activated by the High capability designation. Data processing agreements should be reviewed to confirm that routing signal and interaction data handling is covered for developer-specific deployments.
This provision defines the model variants available to API developers, which determines the capability ceiling for applications built on the GPT-5 platform and establishes that the High capability bio/chem classification and safe-completions restrictions apply to the thinking model variants accessible via the API.
This provision establishes the specific model variants available to developers through the API, which governs the capabilities and associated safeguards, including safe-completions and the High bio/chem capability classification, that apply to applications built on gpt-5-thinking variants.
No. ConductAtlas is an independent monitoring service. We are not affiliated with, endorsed by, or sponsored by OpenAI.