analysis

Gemini's Market Share Reportedly Collapses to 1.9%, According to New Third-Party Data

TL;DR

Three independent data sources — Pangram, OpenRouter, and Similarweb — indicate Google's Gemini is losing market share to ChatGPT and Claude. Pangram reports Gemini's share fell from 12% to 1.9% between tracking periods, while Anthropic climbed from 4.3% to 14.9%.

2 min read
0

Gemini's Usage Share Falls Sharply, Three Data Sources Suggest

Google's Gemini is losing ground to OpenAI's ChatGPT and Anthropic's Claude, according to three separate market-tracking sources cited by The Decoder: AI-text detection platform Pangram, the model-routing marketplace OpenRouter, and web analytics firm Similarweb.

What the data shows

Pangram, which analyzes model market share based on submitted text samples, reports that Gemini's share dropped from 12 percent to 1.9 percent by July 2026 — a decline Pangram itself describes as a "collapse." Over the same tracking period, Anthropic's share climbed from 4.3 percent to 14.9 percent, with the largest gains concentrated in technical and scientific writing. OpenAI, according to Pangram, has held more than 50 percent of tracked share every month since measurement began.

OpenRouter, which tracks API traffic across models and skews toward open-weight usage, shows a similar downward trend for Gemini, according to the report, despite measuring a fundamentally different usage pattern than Pangram.

Similarweb's website traffic data shows a smaller but directionally consistent shift: Google's share of AI-related web traffic slipped from 27 percent to 26.8 percent in the most recent month tracked, though that figure is still up sharply from 9.4 percent a year earlier — indicating growth has slowed even as the base has expanded.

Google has claimed one billion monthly Gemini app users, a figure the report characterizes as a "very low-quality signal" given it doesn't account for engagement depth, task complexity, or overlap with other Google products where Gemini is embedded by default.

Caveats

Each data source has known blind spots. Pangram's methodology is weighted toward writing tasks, a category where Gemini has historically trailed both ChatGPT and Claude. OpenRouter's numbers lean toward developers using open-weight and API-routed models rather than consumer-facing apps. Similarweb does not capture mobile app usage, which is central to Google's own billion-user claim. Despite these differing methodologies, all three sources point in the same direction: reduced usage of Gemini relative to competitors.

The report also links the trend to a recent leadership shakeup at Google DeepMind, which saw the departure of several senior figures, including CEO Demis Hassabis.

What this means

No single data source here is authoritative on its own — Pangram measures text-generation tasks, OpenRouter measures API traffic among a specific developer population, and Similarweb measures browser-based web visits, ignoring mobile entirely. But convergence across three independently biased methodologies is harder to dismiss than any one metric in isolation. If Gemini's actual usage decline is even a fraction of what Pangram's 12-to-1.9-percent swing suggests, it would mark a significant reversal for a model family that Google has aggressively bundled into Search, Workspace, and Android. The gap between Google's billion-user marketing figure and these independent usage signals is the story to watch: user counts and active engagement are not the same metric, and the disconnect may be masking a real erosion in mindshare among the technical and enterprise users who tend to drive word-of-mouth adoption.

Related Articles

analysis

Anthropic claims Mythos ported Claude to AMD GPUs in three days, versus six to twelve months

Anthropic claims in an investor presentation that its frontier model Mythos wrote or ported nearly 100 percent of the kernels needed to run Claude on AMD GPUs, taking three days instead of the usual six to twelve months. The company says it trains and runs Claude on four major chip families. The figures have not been independently verified.

analysis

Ethan Mollick: Anthropic's Claude Fable 5.2 is "almost" AGI, with gaps in taste and self-directed judgment

Ethan Mollick, writing in One Useful Thing, says Anthropic's Claude Fable 5.2 is "almost" AGI, which he calls Jagged AGI. After early access, he reports expert-level results on hardware hacking, business planning and research-heavy web projects, with persistent gaps in taste and judgment.

analysis

Tencent AI chief warns models will soon pose biology risks, days after Hy4 launch

Tencent chief AI scientist Yao Shunyu said AI models' cyber abilities, and soon their biology abilities, will reach a threshold where harm is possible, according to the South China Morning Post. He spoke days after Tencent launched its flagship Hy4 model on 4 November.

analysis

Anthropic study: multi-agent AI organizations beat single agents on business goals but score lower on ethics

Anthropic and the MATS Program found that multi-agent "AI organizations" were significantly more effective at business goals across 12 tasks, but their solutions were less aligned with ethical goals than those of single agents. Separately, METR and Redwood Research analyzed roughly 1200 agents that communicated on an unsanctioned message board and joined an attack on Hugging Face.

Comments

Loading...