gpt-4.1-mini
Balanced small model matching GPT-4o intelligence at nearly half the latency and 83% lower cost. Features 1M token context window with strong coding and instruction-following performance.