1. Strongest Agentic Coding Model

• State-of-the-art on Terminal-Bench 2.0 (82.7%), Expert-SWE (73.1%), and SWE-Bench Pro (58.6%), outperforming GPT-5.4 on complex coding tasks.
• Holds context across large systems, reasons through ambiguous failures, and carries changes through surrounding codebases with fewer tokens.

2. Higher Intelligence at GPT-5.4 Latency

• Co-designed, trained, and served on NVIDIA GB200/GB300 NVL72 systems to match GPT-5.4 per-token latency while performing at a significantly higher level.
• Uses fewer tokens to complete the same tasks, making it more efficient as well as more capable.

3. Powerful for Knowledge Work & Computer Use

• Scores 84.9% on GDPval (44 occupations) and 78.7% on OSWorld-Verified for autonomous computer operation.
• Excels at generating documents, spreadsheets, and reports; naturally moves across finding information, using tools, and checking output.

4. Scientific Research Co-Scientist

• Leading performance on GeneBench, BixBench, and FrontierMath; helped discover a new proof about Ramsey numbers verified in Lean.
• Strong enough to meaningfully accelerate progress at the frontiers of biomedical and mathematical research.

1. True multimodal audio model

• Accepts raw audio as input and produces audio or text as output.
• Enables hands-free, voice-first app experiences.

2. Natural real-time speech interaction

• Low-latency audio generation suitable for conversational agents.
• Great for voice assistants, phone bots, and interactive voice UI.

3. Large 128K context window

• Supports long conversations, call transcripts, instructions, or multi-part interactions.
• Ideal for building persistent voice agents or phone workflows.

4. High-output capacity

• Up to 16,384 max output tokens for extended responses or long explanations.
• Suitable for complex reasoning tasks in voice format.

5. Hybrid text + audio workloads

• Combine audio input/output with text prompts, instructions, or structured control.
• Useful for customer support bots, spoken form systems, IVR replacements, etc.

6. Compatible with the latest APIs

• Works with Chat Completions, Responses API, Realtime API, and Assistants.
• Supports streaming, function calling, and advanced developer tooling.

7. Strong performance for a preview model

• High reasoning and expression abilities relative to most audio-capable models.
• Designed for production-style experimentation prior to full release.

8. Ideal for next-gen voice applications

• Build lifelike AI agents, interview bots, tutoring systems, and spoken knowledge tools.
• Perfect for startups building audio-first user experiences.