AI is changing the world. Don't stay behind. Clear summaries, community insight, delivered without the noise. Subscribe to never miss a beat.

Privacy

Contact

Back to all news

gpt-5 openai ai-agents code-generation

GPT-5.4

openai.com

March 5, 2026

17 min read

🔥🔥🔥🔥🔥

75/100

Summary

GPT-5.4, now available in ChatGPT and the API, is the most capable and efficient frontier model designed for professional work. It combines advanced reasoning, coding capabilities from GPT-5.3-Codex, and enhanced agentic workflows for improved performance on complex tasks.

Key Takeaways

OpenAI released GPT-5.4, the most capable and efficient model for professional work, enhancing reasoning, coding, and agentic workflows.
GPT-5.4 achieves a new state of the art on the GDPval benchmark, matching or exceeding industry professionals in 83.0% of comparisons.
The model supports up to 1 million tokens of context, allowing for complex workflows and improved task execution across applications.
GPT-5.4 shows significant improvements in creating and editing spreadsheets and presentations, with mean scores of 87.3% and 68.0% preference from human raters, respectively.

Read original article

Community Sentiment

Mixed

Positives

GPT-5.4's 1M context window significantly exceeds the ~200k tokens supported by other models, potentially enabling more complex interactions and improved performance in extensive tasks.
The pricing for GPT-5.4 is competitive, being much cheaper than Opus 4.6, which could make advanced AI more accessible for developers.

Concerns

The reliance on coordinate-based clicking and screenshot interpretation in GPT-5.4 raises concerns about efficiency and practicality compared to using APIs directly.
The confusion around versioning and model availability suggests a lack of clarity and stability in OpenAI's offerings, which could frustrate developers.
Skepticism remains regarding the actual benefits of the 1M context window, as previous models like Codex and Opus have shown weaknesses despite their capabilities.