OpenAI launches GPT-4.1 to improve software engineering

By Canuto

OpenAI launches the GPT-4.1 model, focused on coding, competing with rivals such as Google and Anthropic in the race to develop sophisticated programming models.
***

• GPT-4.1 of OpenAI focuses on advanced coding.
• Multimodal models with context window of 1 million tokens.
• Competence between OpenAi, Google and Anthropic in programming.

---

OpenAI has presented its new GPT-4.1 model, specifically designed to improve coding skills.

With different versions such as GPT-4.1 Mini and Nano, these models promise to optimize the coding and monitoring of instructions, offering a context window of up to 1 million tokens.

This allows the model to process approximately 750,000 words simultaneously, an unprecedented capacity in artificial intelligence focused on programming.

In a context where technological giants such as Google and Anthropic accelerate their efforts to develop advanced programming models, Openai has added a new competitive element to the market.

According to TechCrunch, the GEMINI 2.5 PRO of Google and the Claude 3.7 Sonnet of Anthropic already offer similar functions with impressive results in coding tests.

Growth and competition in artificial intelligence

Openai’s ultimate goal goes beyond simple improvements; The company seeks to create a ‘agentico software engine’.

As the CFO Sarah Friar said at a technological summit in London, this model will not only schedule complete applications but will also be responsible for errors testing, documentation elaboration and quality assurance.

The recent GPT-4.1 launch represents a significant advance towards this goal, since it exceeds its predecessors in specific coding tests such as Swe-Bench.

Despite its improved capacity, the initial model still fights with certain tasks, a common reality in the current AI, where even the most advanced models sometimes fail to solve more complex problems without introducing additional errors.

Economy and efficiency of GPT-4.1

The GPT-4 model not only offers technological advances, but also a more competitive economic adjustment compared to others.

The cost is USD $ 2 per million input tokens and USD $ 8 per million output tokens.

Mini and Nano versions, although less precise, offer faster speeds and reduced prices, which increases accessibility for developers with different budgetary needs.

These economic and technical characteristics position OpenAI models as viable and attractive options in the current market, expanding their possibilities of integration in software engineering applications in the real world.

Limitations and perspectives

Despite its potential, GPT-4.1 is not exempt from limitations.

Openai acknowledges that The model becomes less reliable with the increase in the number of input tokensfalling its accuracy of 84% to a mere 50% in certain internal tests.

The company, however, considers these development phases as necessary steps towards perfection once they address these challenges.

With the promise that future models will address these limitations, Openai continues in the technologically competitive career against Google, Anthropic and other key players who seek to define the future of AI in programming.

---

Original image of Diariobitcoin, created with artificial intelligence, for free use, licensed under public domain.
This article was written by an AI content editor and reviewed by a human editor to guarantee quality and precision.

WARNING: Diariobitcoin offers informative and educational content on various topics, including cryptocurrencies, AI, technology and regulations. We do not provide financial advice. Cryptactive investments are high risk and may not be adequate for all. Investigate, consult an expert and verify the applicable legislation before investing. I could lose all its capital.