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. Extremely low-cost text model

• One of the cheapest legacy models available.
• Suitable for very high-volume workloads with simple requirements.

2. Good for lightweight NLP tasks

• Classification, summarization, rewriting, paraphrasing, intent detection.
• Works for simple logic tasks and short reasoning sequences.

3. Works well for basic chatbots

• Optimized for Chat Completions API, originally powering early ChatGPT use cases.
• Good for rule-based or templated conversation flows.

4. Stable and predictable outputs

• Legacy behavior makes it suitable for systems built years ago that rely on its quirks.
• Good for backward compatibility or long-term enterprise pipelines.

5. Supports fine-tuning

• Useful for teams maintaining older fine-tuned GPT-3.5 models.
• Allows domain-specific compression of older datasets.

6. Limited capabilities compared to newer models

• No vision, no audio, no streaming, and no function calling.
• Much weaker reasoning and correctness vs GPT-4o mini or GPT-5.1.

7. Small context window (16K)

• Limited for multi-document tasks or long conversations.
• Best used for short, simple prompts or structured tasks.

8. Recommended migration path

• OpenAI explicitly recommends using GPT-4o mini instead.
• 4o mini is cheaper, smarter, faster, multimodal, and far more capable.