Pull to refresh
Logo
Daily Brief
Following Why Ranks Sign Up
OpenAI makes GPT-5.5 Instant ChatGPT's new default model

OpenAI makes GPT-5.5 Instant ChatGPT's new default model

New Capabilities

More accurate on high-stakes questions, more personalized, less cluttered—and arriving as ChatGPT's lead over rivals narrows

May 5th, 2026: GPT-5.5 Instant becomes default

Overview

ChatGPT receives roughly a billion visits a month, and on May 5 those visitors began talking to a different model by default. OpenAI replaced GPT-5.3 Instant—the everyday workhorse it shipped earlier this year—with GPT-5.5 Instant, a faster system the company says produces 52.5% fewer hallucinated claims on medical, legal, and financial questions. The new model can also pull context from a user's past chats, uploaded files, and Gmail to personalize answers, and it appears in OpenAI's developer interface as 'chat-latest.'

The change lands in a market OpenAI no longer dominates the way it once did. ChatGPT's share of AI-chatbot web traffic has fallen from roughly 87% to about 68% over the past year, with Google's Gemini climbing to about 18% and Anthropic's Claude winning a disproportionate share of enterprise deals. Upgrading the default model is OpenAI's most direct lever for keeping casual users—the population most likely to drift when answers feel cluttered, slow, or wrong.

Why it matters

ChatGPT's default model is the AI most people actually use—and it's now meaningfully more accurate on the questions where wrong answers cost the most.

Play on this story Voices Debate Predict

Key Indicators

52.5%
Fewer hallucinated claims
OpenAI's measured reduction versus GPT-5.3 Instant on high-stakes medical, legal, and financial prompts.
37.3%
Fewer inaccurate claims
Reduction on the harder real-world conversations users had previously flagged for factual errors.
81.2
AIME 2025 math score
Up from 65.4 for GPT-5.3 Instant on the standardized math benchmark.
68%
ChatGPT market share
Down from 87% a year earlier as Gemini and Claude absorbed traffic.
3 months
GPT-5.3 Instant grace period
How long paid subscribers can keep using the previous default before it is retired.

Interactive

Exploring all sides of a story is often best achieved with Play.

Ever wondered what historical figures would say about today's headlines?

Sign up to generate historical perspectives on this story.

Log in Sign Up

People Involved

Sam Altman
Sam Altman
Leading OpenAI's accelerated model-iteration strategy

Organizations Involved

OpenAI
OpenAI
AI model developer
Shipping a new default model to defend ChatGPT's lead

The company behind ChatGPT and the GPT model family.
Anthropic
Anthropic
AI model developer
Primary competitor; gaining ground in enterprise

Maker of the Claude family of AI assistants and OpenAI's most direct rival.
Google DeepMind
Google DeepMind
Technology platform; AI assistant operator
Fastest-growing rival to ChatGPT in consumer share

Owner of Gemini, the AI assistant integrated across Google Search, Workspace, and Android.

Timeline

  1. GPT-5.5 Instant becomes default

    Product Release

    OpenAI makes GPT-5.5 Instant the default in ChatGPT and ships it via the API as 'chat-latest,' citing 52.5% fewer hallucinations on high-stakes prompts and new memory features that read past chats, files, and Gmail.

  2. GPT-5.5 foundation model released

    Product Release

    OpenAI introduces GPT-5.5 to Plus, Pro, Business, and Enterprise tiers, with GPT-5.5 Pro restricted to higher tiers.

  3. GPT-5.4 launches

    Product Release

    OpenAI ships GPT-5.4 Thinking and Pro variants; mini and nano follow on March 17.

  4. Older models retired

    Product Change

    GPT-4o, GPT-4.1, o4-mini, and the original GPT-5 are removed from ChatGPT; users petition to keep GPT-4o, calling it a 'best friend.'

  5. GPT-5 family becomes default

    Product Release

    OpenAI replaces GPT-4o with the GPT-5 family as ChatGPT's default model, prompting user complaints about lost personality.

  6. ChatGPT launches

    Product Release

    OpenAI releases ChatGPT as a free research preview; it becomes the fastest-growing consumer software product on record.

Scenarios

Predict which scenario wins. Contrarian picks score more — points lock in when the scenario resolves.

Log in to predict. Track your picks, climb the leaderboard. Log in Sign Up
1

Accuracy gains slow ChatGPT's market-share slide

Reduced hallucinations on medical, legal, and financial questions are the most cited reason users say they switched away from ChatGPT. If GPT-5.5 Instant's claimed 52.5% improvement holds up under independent testing, OpenAI may stabilize its share around the high-60% range as casual users find fewer reasons to try Gemini or Claude for serious questions.

Discussed by: TechCrunch, SiliconANGLE, The Decoder
Consensus —
2

Memory features trigger privacy backlash

Granting the default model access to a user's Gmail, files, and past conversations creates a much larger personal-data surface than prior versions. A high-profile leak, accidental cross-conversation disclosure, or regulatory inquiry in the European Union could force OpenAI to restrict the personalization features it just rolled out, blunting their competitive value.

Discussed by: 9to5Mac, MacRumors, privacy researchers
Consensus —
3

Anthropic answers with a Claude release that widens its enterprise lead

Claude already wins roughly 70% of head-to-head enterprise deals despite a small consumer footprint. A near-term Anthropic release focused on coding, agentic reliability, or cybersecurity—areas where Claude is already preferred—could keep enterprise revenue migrating to Anthropic even as OpenAI tightens its consumer position.

Discussed by: The Register, AI Business Weekly, enterprise analysts
Consensus —
4

Google closes the personalization gap by leveraging Workspace

OpenAI's new memory features replicate, in part, what Google can do natively across Gmail, Docs, and Calendar. If Google ships a Gemini upgrade that more deeply integrates Workspace data, the personalization advantage OpenAI just claimed could shrink within months, leaving improved accuracy as the main differentiator.

Discussed by: Axios, Similarweb analysts
Consensus —

Historical Context

Internet Explorer loses default-browser dominance to Chrome (2008–2014)

September 2008 – May 2014

What Happened

Microsoft's Internet Explorer held more than 70% of web-browser market share when Google launched Chrome in September 2008. Chrome shipped faster release cycles, better JavaScript performance, and tighter integration with Google services. By May 2014, Chrome had passed IE in worldwide usage.

Outcome

Short Term

Microsoft accelerated IE updates and shipped IE9 and IE10 with major performance improvements, but the version-bump cadence could not match Chrome's silent auto-updates.

Long Term

IE was eventually replaced by Edge, and Microsoft adopted Chrome's underlying Chromium engine in 2020, conceding the platform layer to Google.

Why It's Relevant Today

OpenAI's faster default-model cadence—and the strategic choice to give the default direct access to Gmail-style personal context—mirrors Microsoft's defensive moves. It illustrates how a default product can lose share gradually even when each new version is genuinely better than the last.

Google's 'Panda' and 'Penguin' search updates (2011–2012)

February 2011 – April 2012

What Happened

Google rolled out major search-ranking updates—Panda in February 2011 and Penguin in April 2012—aimed at reducing low-quality and manipulative content in results. The changes affected an estimated 12% of English-language queries and reshuffled traffic for entire categories of websites.

Outcome

Short Term

Search-quality complaints fell and content farms lost rankings; affected publishers reported sharp traffic drops within weeks.

Long Term

Google's iteration on the default-ranking algorithm became the central mechanism through which it defended search dominance against Bing and, later, AI-powered alternatives.

Why It's Relevant Today

Like Google's algorithm, ChatGPT's default model is the product for the vast majority of users. Iterating on the default—rather than launching new SKUs—is the highest-leverage move OpenAI can make to defend share.

Apple Maps replaces Google Maps as iOS default (2012)

September 2012

What Happened

Apple replaced Google Maps with its own Maps app as the default in iOS 6, instantly putting Apple Maps on hundreds of millions of iPhones. Early versions had visible accuracy problems—missing landmarks, mislabeled towns, distorted satellite imagery—prompting a public apology from CEO Tim Cook and the firing of the executive responsible.

Outcome

Short Term

Users complained loudly; Google released a standalone Google Maps app for iOS that was downloaded 10 million times in 48 hours.

Long Term

Apple invested heavily in mapping data and rebuilt the product over several years; by the early 2020s, Apple Maps had reached parity for most use cases on iOS.

Why It's Relevant Today

Replacing a default that hundreds of millions of people rely on is unforgiving when accuracy slips. OpenAI's hallucination-reduction claim matters precisely because the default model touches every casual ChatGPT session, and visible regressions in medical, legal, or financial answers would land the way Apple Maps' missing landmarks did.

Sources

(10)
+4
OpenAI TechCrunch 9to5Mac OpenAI Help Center Axios SiliconANGLE The Decoder Similarweb / Vertu OpenAI OpenAI Help Center