OpenAI is attempting to demonstrate its commercial viability to capital markets through a price increase for GPT-5.4.

On March 6, Beijing time, OpenAI launched GPT-5.4, its latest-generation AI model. OpenAI has defined this new model as the company's current "most authentic and efficient" model—consuming fewer tokens and operating at faster speeds. Furthermore, within ChatGPT, GPT-5.4 exhibits enhanced reasoning capabilities for deep web research, maintaining context better during longer thinking periods, and allowing users to interrupt the model mid-response to add instructions or change direction. Next, OpenAI will roll out GPT-5.4 Pro in ChatGPT and its API, targeting users who require the highest performance for complex tasks. According to disclosures, GPT-5.4 integrates OpenAI's recent advancements in reasoning, coding, and Agent workflows. It combines the industry-leading coding capabilities of GPT-5.3-Codex while improving the model's performance across various tools, software environments, and specialized tasks involving spreadsheets, presentations, and documents. The result is a model capable of accurately, effectively, and efficiently completing complex real-world tasks—delivering what users request with less back-and-forth communication. Additionally, in ChatGPT, GPT-5.4's reasoning mode can now provide its thinking plan in advance, allowing users to adjust its direction at any point during its response. The model's reasoning ability has also been enhanced, showing particular strength in deep web research, especially for highly specialized queries. It also maintains contextual coherence more effectively for problems requiring extended thought. These improvements collectively lead to higher-quality, faster responses that remain tightly focused on the task at hand. Notably, incorporating these capabilities, GPT-5.4 has become the first general model within OpenAI to natively include Computer Use functionality. Previously, related capabilities like ChatGPT Atlas and AutoGPT-style Agents were add-ons/plugins/independent agents, not native model features. Prior to GPT-5.4, products like Anthropic's Claude Opus 4.6 and Google's Gemini Advanced had already captured developer attention in the Computer Use domain. Although OpenAI emphasizes that the new model consumes fewer tokens and runs faster, the performance improvements come with an overall price hike. Official disclosures state that GPT-5.4 input pricing has risen to $2.50 per million tokens, with output pricing at $15 per million tokens. The Pro version's input pricing is as high as $30 per million tokens, with output pricing at $180 per million tokens. Regarding the magnitude of this price increase, some online commentators believe the price is acceptable if it aligns with user ROI. However, others have reported that a simple greeting to GPT-5.4 cost them $80. "Everyone is saying GPT-5.4 is at the level of AGI (Artificial General Intelligence), the smartest model, but do you have AGI-level questions to ask?" one user questioned. Behind the performance improvements and price increases for GPT-5.4 lies OpenAI's revenue pressure within the current AGI competitive landscape. Although the company recently announced the completion of a new $110 billion funding round, Nvidia CEO Jensen Huang, representing a key OpenAI investor, indicated that this $30 billion investment might be the last. The likelihood of realizing the $100 billion supercomputing facility agreement reached last year is also minimal. The core reason is OpenAI's impending IPO, which will change its capital structure and financing strategy, making it difficult for Nvidia to continue making large-scale investments as a private equity investor. Furthermore, reports suggest that this substantial $110 billion funding round is not a lump-sum cash injection but rather a contingent direct agreement. Coupled with the competitive pressure from companies like Anthropic and Google in the AI space, OpenAI is not only adjusting its cost structure to be more pragmatic but is also exploring multiple avenues to expand its revenue streams. This includes recently securing a US Department of Defense order that Anthropic had lost. The price hike for GPT-5.4 represents a test for OpenAI on the eve of its IPO. The company is attempting to prove to investors and capital markets that its product's performance is so high that businesses and users are willing to pay the increased price.