Gemini Market Share Falls as ChatGPT & Claude Gain Ground

Google’s Gemini is losing ground to ChatGPT and Claude across several measures of AI usage, according to new market data, raising fresh questions about the company’s position in an increasingly competitive generative AI market. Data from AI text-analysis platform Pangram shows Gemini’s share of submitted texts falling sharply in July, while OpenAI maintained a dominant position and Anthropic’s Claude continued to gain ground.

The findings come despite Google reporting that the standalone Gemini app has reached 1 billion monthly active users. The contrasting numbers highlight the difficulty of measuring the AI market through a single metric: user counts, web traffic, model usage and enterprise adoption can tell very different stories about which AI platforms are actually gaining momentum.

Gemini’s Market Share Falls Sharply

Pangram, an AI text-recognition platform, has been tracking which AI models are responsible for submitted text. Its latest figures show a significant change in the competitive landscape.

OpenAI has maintained more than 50% of the tracked market every month since Pangram began monitoring it. Anthropic’s share increased from 4.3% to 14.9%, with particularly strong usage in technical and scientific fields.

Gemini moved in the opposite direction. Its share fell from 12% to just 1.9% in July 2026, a decline Pangram described as a collapse.

AI Model	Earlier Share	July 2026 Share	Trend
OpenAI models	Above 50%	Above 50%	Dominant
Anthropic	4.3%	14.9%	Rising
Google Gemini	12%	1.9%	Falling sharply

The data does not represent the entire AI market, but the scale of Gemini’s decline is notable because it suggests that model usage can diverge significantly from headline user numbers.

OpenRouter Data Also Shows Gemini Losing Ground

Pangram is not the only source pointing toward weaker Gemini usage.

The Decoder reported that OpenRouter data, which tracks API usage across a large selection of AI models, also showed Gemini losing market share. OpenRouter and Pangram measure different types of activity, but both datasets point in the same broad direction: developers and AI users are increasingly allocating workloads to competing models.

This distinction is important.

A consumer may have access to Gemini through Google’s products without regularly using Gemini’s underlying models for demanding tasks. Developers, businesses and technical users may instead select models based on factors such as reasoning quality, coding performance, cost and reliability.

That means Google’s enormous distribution advantage does not necessarily translate into equivalent model usage.

Claude Is Gaining Among Technical Users

Anthropic’s Claude has emerged as one of the clearest beneficiaries of the changing market.

Pangram’s data shows Claude’s share rising from 4.3% to 14.9%. Much of the increase is concentrated in technical and scientific use cases, areas where developers and researchers often place greater emphasis on reasoning, coding and long-context capabilities.

Claude’s enterprise momentum also provides another indication of its growing importance.

An earlier survey of 100 CIOs from large companies found that OpenAI remained the leading provider in production, with 78% of surveyed companies using its models. However, Anthropic had increased its enterprise penetration to 44%, a 25% increase from May 2025.

This suggests that Anthropic is building strength in areas where companies use AI for professional and technical workloads.

ChatGPT Remains the Market Leader

Despite growing competition, ChatGPT continues to hold the strongest overall position.

Similarweb data published earlier in 2026 showed ChatGPT accounting for 61.7% of global AI web traffic in February, down from 75.7% a year earlier. Gemini had increased its share dramatically during that period, rising from 5.7% to 24.4%.

That earlier trend is important because it demonstrates how quickly market shares can change.

The latest Pangram data tells a different story from the Similarweb numbers, with Gemini now showing significant weakness in model usage. The difference illustrates why AI market-share estimates should not be treated as a single definitive leaderboard.

Web visits, API calls, model selection and app usage measure different parts of the market.

Google Still Has a Huge User Base

Gemini’s reported 1 billion monthly users complicate the narrative that Google is simply losing the AI race.

Google CEO Sundar Pichai said in August that 1 billion people were using Gemini each month, making it Google’s fastest-growing product. OpenAI also recently disclosed that ChatGPT had crossed the 1 billion monthly-user milestone.

The two companies therefore both have enormous consumer reach.

However, Google’s figure refers specifically to the Gemini app and does not necessarily mean that all users are actively relying on Gemini for complex AI tasks every month.

Google also benefits from its enormous Android ecosystem and the ability to integrate Gemini across Search and other products.

Web Traffic Tells a Different Story

Similarweb’s data has previously shown Gemini gaining significant ground against ChatGPT.

In May 2026, Similarweb data showed ChatGPT’s share of AI website traffic falling to 53.7%, compared with 77.6% a year earlier. Gemini increased from 7.3% to 26.7% over the same period, while Claude also gained share, approaching 8%.

This demonstrates why Gemini’s current weakness in some model-usage data should not be interpreted as a universal collapse.

Google can drive Gemini usage through Android, Search and other products in ways that are not fully captured by model-level datasets.

The company’s distribution advantage remains one of its biggest strengths.

Why Gemini May Be Losing Technical Workloads

One possible explanation for the latest data is the growing preference among technical users for models from OpenAI and Anthropic.

Developers often evaluate AI systems based on coding performance, reasoning, context handling and reliability rather than simply the availability of a chatbot.

Anthropic has built a particularly strong position in coding through Claude Code, while OpenAI has expanded ChatGPT into a broader platform with tools, agents and integrations.

Google remains a major AI research organization, but delays and competitive pressure around its flagship models have created concerns about whether Gemini can maintain leadership at the frontier.

Reuters reported in August that Google had delayed a new flagship Gemini version after internal testing showed its performance was still lagging rivals in areas including coding. The company has also undergone a major leadership restructuring within Google DeepMind

Google Is Responding With a Broader AI Strategy

Google is not relying exclusively on the Gemini chatbot.

The company has integrated Gemini deeply across its consumer and enterprise ecosystem, including Search, Android, Workspace and other products.

Google is also developing specialized models and open-source systems through its Gemma family.

Earlier this year, Google launched Gemma 4, an open-source model designed to run on devices while supporting text, images and audio. The company said the Gemma family had accumulated more than 400 million downloads since its first generation.

This broader strategy gives Google several ways to distribute AI even if standalone Gemini model usage fluctuates.

The AI Market Is Becoming More Fragmented

The latest data also points to a broader change in the AI industry.

The market is no longer dominated by a single chatbot in every category. ChatGPT remains strongest overall, Claude is gaining among enterprise and technical users, while Gemini has a major advantage through Google’s consumer ecosystem.

An earlier a16z analysis found that about 20% of weekly ChatGPT web users also used Gemini during a given week, indicating that consumers increasingly use multiple AI assistants rather than committing exclusively to one platform. The same research found that 81% of surveyed companies used three or more model families.

This multi-model behaviour could become increasingly important.

Instead of choosing one AI provider, companies may use different models for coding, research, customer service, document analysis and other tasks.

User Numbers May Become Less Important

The competing data also suggests that headline user numbers may not be the best measure of AI market leadership.

Google’s 1 billion Gemini users represent enormous reach, but the number does not reveal how frequently those users interact with the service, what tasks they perform or which models handle their most valuable workloads.

Similarly, web traffic does not capture API usage, enterprise deployments or AI coding tools.

The industry may therefore move toward more sophisticated measures of market share that combine consumer engagement, enterprise usage, API consumption, revenue and developer activity.

What the Shift Means for Google

For Google, the challenge is not simply attracting users to Gemini.

The company needs those users to regularly choose Gemini for increasingly important AI tasks.

That means Gemini must remain competitive on model quality while Google continues to improve product integration, pricing and developer tools.

The company’s advantage is its enormous ecosystem. Its weakness is that competitors can move quickly and focus their resources on specific AI use cases.

Anthropic, for example, has concentrated heavily on enterprise and coding applications, while OpenAI continues expanding ChatGPT into a broader consumer and productivity platform.

Google therefore has to compete across multiple fronts simultaneously.

Looking Ahead

The latest Pangram and OpenRouter data suggests that Gemini is losing ground in some areas of AI model usage, particularly compared with the momentum being recorded by OpenAI and Anthropic. Gemini’s reported decline from 12% to 1.9% in Pangram’s tracked usage is significant, while Claude’s rise from 4.3% to 14.9% highlights the growing strength of Anthropic among technical and scientific users

However, Google’s reported 1 billion monthly Gemini users and its powerful Android, Search and Workspace distribution mean that declaring Gemini a loser would be premature. The AI market is increasingly fragmented, with different platforms leading in different categories. The bigger test for Google will be whether it can convert its enormous distribution advantage into sustained model usage, enterprise adoption and revenue as ChatGPT and Claude continue to compete aggressively for the most valuable AI workloads.