GPT-5.4 Defines OpenAI’s Current Model Lineup

OpenAI moved into April 2026 with strong momentum. The company had just closed a record $122 billion funding round that valued it at $852 billion, and it was also only about a month past the release of GPT-5.4, described as its most capable model so far. At the same time, attention has started shifting toward what may come next, with growing discussion around GPT-5.5. Even so, no private beta has been confirmed, and the immediate priority appears to be tightening the current model lineup while setting the stage for a potentially larger release ahead.

GPT-5.4 release and architecture

GPT-5.4 launched on March 5 and expanded on the GPT-5 architecture that first arrived in August 2025. The model was released in Standard, Thinking, and Pro versions, giving OpenAI a broader structure across different use cases and performance needs.

GPT-5.4 capabilities and benchmark performance

One of the biggest additions in GPT-5.4 was native computer-use functionality. On the OSWorld-Verified benchmark, it scored 75%, which placed it above human performance at 72.4%. It also reached a record 83% on OpenAI’s GDPval test for knowledge work tasks.

That positioning reinforces GPT-5.4 as the center of OpenAI’s current offering. The emphasis is not just on generating outputs, but on handling more demanding work with stronger reasoning and task execution.

Smaller GPT-5.4 variants

OpenAI followed the main launch with GPT-5.4 mini and nano on March 17. Those additions extended the family further and showed that the company was continuing to build around the same architecture rather than immediately shifting public attention to an officially announced successor.

GPT-4o Retirement Marks the End of an Earlier Model Era

OpenAI has now fully retired GPT-4o from all plans. That move closes the chapter on the model family that came before GPT-5 and signals a cleaner transition to the newer generation.

What GPT-4o retirement means

Retiring GPT-4o from all plans suggests OpenAI is simplifying its active lineup. Instead of maintaining overlap between generations, the company appears to be concentrating attention on GPT-5-based models.

This also sharpens the contrast between the company’s present focus and the market’s future expectations. GPT-4o is out, GPT-5.4 is established, and speculation about GPT-5.5 is filling the gap between what is available now and what users expect could come next.

GPT-5.5 Speculation Continues to Grow

Discussion around GPT-5.5 has intensified, but official confirmation is still missing. That gap between public curiosity and company silence is a big part of why anticipation keeps building.

Reported codename and pretraining status

The next major model has reportedly carried the codename “Spud.” According to prediction market reports on Polymarket, pretraining has finished, and a rollout is expected in Q2 2026. Before that, an earlier internal codename, “Garlic,” surfaced in January.

Reports tied to that earlier codename suggested the model performs better than competitors on logic and coding tasks. Still, OpenAI has not officially announced the model.

Release timing expectations

Prediction markets on Manifold place the probability of a release before June 2026 at 73%. That does not amount to confirmation, but it shows how strongly expectations have formed around a near-term launch window.

Reports from The Information in early March also said OpenAI’s next model would include “extreme” reasoning capabilities. If that proves accurate, GPT-5.5 may be positioned less as a routine update and more as a notable step forward in advanced reasoning.

Spring Update as a possible announcement stage

OpenAI’s Spring Update, expected in late April, could become the setting for a formal reveal. Right now, that remains a possibility rather than a confirmed plan. But with GPT-5.4 already anchoring the current lineup and GPT-4o now retired, the timing has fueled even more attention around what OpenAI may announce next.

OpenAI’s Broader Expansion Adds to the Momentum

Model speculation is happening alongside a period of rapid company expansion. OpenAI is not operating in a quiet holding pattern. It is moving fast on multiple fronts.

Record funding round and valuation

The company recently completed a $122 billion funding round, reaching a valuation of $852 billion. SoftBank co-led the round along with Andreessen Horowitz, D.E. Shaw Ventures, and Microsoft.

That scale matters because it shows the level of backing behind OpenAI at the same time it is consolidating its model lineup and preparing for future releases.

Workforce growth plans

According to Reuters, OpenAI plans to nearly double its workforce to 8,000 by the end of 2026. That kind of hiring plan points to a company scaling aggressively, not just maintaining its current pace.

TBPN acquisition

On April 2, OpenAI also announced its acquisition of TBPN, a tech-focused talk show. That adds another signal that the company is expanding its reach while public attention around its model roadmap keeps rising.

Why GPT-5.5 Anticipation Feels So Intense Right Now

The current moment combines three things at once: GPT-5.4 is already live, GPT-4o has been fully retired, and OpenAI is in a period of extraordinary financial and organizational expansion. That combination naturally pushes attention toward the next major release.

A lineup in transition

There is a visible transition underway. GPT-4o has exited. GPT-5.4 and its variants now define the active lineup. And the next model is being discussed widely enough that even without official confirmation, expectations are becoming part of the broader conversation.

Consolidation before a bigger release

OpenAI’s near-term focus appears to be consolidation rather than public beta activity for GPT-5.5. No private beta has been confirmed. That matters because it suggests the company may be making sure its existing lineup is fully in place before introducing what could be a more ambitious model.

GPT-5.4 and GPT-5.5 Outlook

GPT-5.4 currently stands as the centerpiece of OpenAI’s model lineup, with Standard, Thinking, Pro, mini, and nano variants establishing the present structure. At the same time, GPT-5.5 speculation keeps accelerating through reported codenames, prediction market signals, and claims around stronger reasoning and coding performance.

What is clear right now is simple: GPT-4o is gone, GPT-5.4 is firmly in place, and OpenAI appears to be preparing carefully for whatever comes next.