OpenAI appears to be preparing for a decisive step in its technological journey, as it nears the unveiling of its latest flagship artificial intelligence model, carrying the internal codename “Spud.” All indicators—ranging from prediction markets and leaked data from testing operations to official statements from top executives—point toward an imminent public rollout that could see the light of day by Thursday, April 23. If this date is confirmed, the release will come at a critical juncture, specifically just weeks after Anthropic’s “Claude Opus 4.7” model successfully reclaimed the lead in complex benchmarks related to coding and logical reasoning. This timing paves the way for a direct and fierce confrontation between two of the world’s leading artificial intelligence developers.
Analysis of prediction markets and live testing
Prediction markets, specifically the Polymarket platform, show a strong trend toward the approaching launch date, with traders now assigning an approximately 80 percent probability to the release of the “GPT-5.5” model on April 23. In contrast, the probability of no new model being released by the end of April has dropped to only about 12 percent. These expectations underwent a dramatic shift and a noticeable surge after API monitoring systems detected activity from the “Spud” model in a large-scale live testing environment on April 19, without any prior announcement or formal preamble from OpenAI.
These speculations were further reinforced by an internal memo sent by Sam Altman to company employees on March 24. Altman confirmed in his memo the completion of the model’s pre-training phase at the company’s “Stargate” facility. He added in his encouraging message that things are moving at a faster pace than anyone expected, heralding that the new release is coming within a few weeks. This timeline aligns perfectly with current expectations, placing the usual four-week post-training period—typically dedicated to safety evaluations, security assessments, and penetration testing—to conclude on April 21, making a launch late in that same week entirely logical.
Competitive response and leadership challenges
This accelerated timeline appears to be no coincidence, but rather a calculated strategic response to fierce market competition. Anthropic launched its advanced “Claude Opus 4.7” model on April 16, which immediately captured first place on the certified software evaluation platform—a global standard for assessing coding capabilities—achieving an exceptional score of 87.6 percent. Meanwhile, OpenAI’s current flagship model, “GPT-5.4,” still lags behind in this vital metric, placing tremendous pressure on the company to reclaim its position as the provider of the best smart models.
Leadership vision and anticipated revolutionary features
Greg Brockman, president of OpenAI, provided the most prominent public comments regarding the new model during an appearance on a prominent tech podcast. Brockman described the “Spud” model as the fruit of two full years of research and development, noting that it carries the characteristics of large-scale models. He emphasized that this release is not merely an incremental improvement like previous updates, but represents a fundamental shift in how the company thinks about model development. This level of enthusiasm and confident language stands out remarkably, especially for a company accustomed to issuing continuous updates, having launched five different fifth-generation variants since August 2025 without generating similar hype.
In a related context, a leaked internal memo prepared by Denise Dresser, the company’s chief revenue officer, revealed other dimensions of this innovative model. Dresser described the model as serving as the solid foundation for an entirely new generation of work and productivity, with an exceptional focus on complex reasoning and advanced capabilities in intent recognition and deep user context comprehension. The memo explained that “Spud” will serve as the primary engine behind the company’s efforts to unify its various platforms, which include the ChatGPT assistant, the Codex autonomous coding agent, and the Atlas browsing agent, integrating them all into a single, unified, and comprehensive platform that delivers a seamless experience for enterprises and individuals.
Competitive landscape and launch roadmap
The “Spud” model enters a crowded competitive arena where performance levels are closer than ever before. While “Claude Opus 4.7” currently sits on the throne of coding, cognitive work, and computer use evaluations, Google DeepMind’s “Gemini 3.1 برو” model stands nearly tied regarding complex scientific reasoning tasks. Despite this, “GPT-5.4” remains the most widely deployed and utilized model in enterprise environments and business sectors. To date, no official evaluations or precise benchmarks have been published for the “Spud” model, and early assessments rely exclusively on leaked outputs from testing phases, which observers have described as representing a qualitative leap and clear progress compared to the previous generation.
The new model is unlikely to carry the name “Spud” upon its commercial launch, as OpenAI has not yet confirmed the final naming convention, while some tech analysts expect it to bear the name “GPT-6,” bypassing sub-naming conventions like “GPT-5.5” to reflect the magnitude of the technological transformation it delivers. Initial access to the new model is expected to be made available exclusively to Plus and Pro tier subscribers on launch day, with the API provided to developers within a few days of that date. As for regular users, the company is likely to release a free, scaled-down version in late May or early June to broaden the user base benefiting from this revolutionary technology.
Frequently Asked Questions
When is OpenAI’s new artificial intelligence model expected to launch?
Prediction markets and leaked data indicate that the company plans to launch the new model, known by the codename “Spud,” on April 23, with direct access provided to Plus and Pro tier subscribers.
What are the most prominent expected improvements in the “Spud” model?
Senior executives have stated that the model represents the culmination of two years of research and development, focusing heavily on complex reasoning capabilities and precise intent recognition, in addition to unifying platforms like ChatGPT, Codex, and Atlas into an integrated interface.
How does this model compete with current market releases?
The launch of this model comes as a direct response to Anthropic’s “Claude Opus 4.7” dominance in coding and reasoning tests, aiming to reclaim technical leadership and surpass major competitors such as Google DeepMind.
