Performance upgrades at Costco prices OpenAI's GPT-4.1 series delivers impressive performance gains, achieving a 54.6% win rate on SWE-bench coding benchmark and outperforming Claude 3.7 Sonnet in real-world tests. The new pricing structure significantly undercuts competitors, with GPT-4.1 costing $2.00 per million tokens for input and $8.00 for output. A 75% caching discount incentivizes prompt reuse, making it particularly attractive for developers. The model offers a full 1M context window as advertised, setting it apart from competitors like Grok, which currently maxes out at 131k tokens.
Performance upgrades at Costco prices OpenAI's GPT-4.1 series delivers impressive performance gains, achieving a 54.6% win rate on SWE-bench coding benchmark and outperforming Claude 3.7 Sonnet in real-world tests. The new pricing structure significantly undercuts competitors, with GPT-4.1 costing $2.00 per million tokens for input and $8.00 for output. A 75% caching discount incentivizes prompt reuse, making it particularly attractive for developers. The model offers a full 1M context window as advertised, setting it apart from competitors like Grok, which currently maxes out at 131k tokens.