GPT-5.5 has barely had time to settle, and people are already talking about what comes next. Over the past 24 hours, a single entry buried in OpenAI's Codex rollout logs has set off a wave of speculation that the company may already be quietly testing GPT-5.6 in the background.

The find came from researcher Haider (@haider1 on X), who was checking internal Codex logs to verify which model was handling requests. Most of the routing pointed to GPT-5.5 as expected, but one entry in the rollout mapping showed something different — a reference to GPT-5.6.

openai is already teasing with gpt-5.6

so if you checked the internal codex logs to verify which model was being used, and you found the rollout mapping

most calls were routed to gpt-5.5, but one entry appears to show gpt-5.6, which means the codex environment may have had… pic.twitter.com/ZjaFCaCIFm

— Haider. (@haider1) April 28, 2026

The post picked up quickly, crossing 200,000 views within hours. But Haider was also the first to pump the brakes.

Shortly after, he noted that the entry had disappeared from session files entirely. His read on it shifted too — from "OpenAI is already teasing GPT-5.6" to something more cautious. "I'd say it's more like a bug or something," he wrote in a follow-up, "because if you check the session files right now, you won't see anything."

So what actually happened? The most likely explanation, based on Haider's own assessment, is that this was canary or eval traffic — a controlled backend test where a small slice of requests gets routed to an experimental build to measure performance before any public rollout. These things happen constantly at large AI labs and rarely signal an imminent release.

That hasn't stopped the conversation from running wild. Some users are reading it as confirmation that OpenAI is in a rapid iteration cycle, pointing to the jump from GPT-5.2 to GPT-5.6 in roughly three to four months. Others are more measured, noting that a single mismatched log entry is thin evidence for anything.

There's a real undercurrent of version fatigue in the replies too. One user summed it up plainly, saying they've stopped chasing intelligence benchmarks and are now focused entirely on reliability. Another went further, saying the community is "tired of the mid-game" and wants GPT-6 already.

OpenAI has not acknowledged the log entry or commented on whether GPT-5.6 exists in any form. The entry itself appears to be gone. What's left is a screenshot, a lot of engagement, and no confirmation from the company either way.