Google disclosed its Gemini models are processing over 10 billion tokens per minute through direct API usage, according to the company's Q4 2025 earnings call. The figure comes as Alphabet posted $113.83 billion in quarterly revenue, up 18% year-over-year, with Google Cloud growing 48% to $17.66 billion.
The token throughput metric is significant for enterprise architecture teams comparing LLM API costs. At that volume, pricing differences between providers compound quickly. Google's Gemini 2.5 Flash targets high-volume batch processing, while Gemini 3 Pro launched in December for complex enterprise workloads. The company's batch API pricing typically undercuts real-time inference for large token volumes, though implementation requires retry logic and timeout handling that real-time APIs abstract away.
Gemini's app user base hit 750 million monthly actives in Q4, up from 650 million in Q3. The ecosystem advantage is clear: integration across Android, Workspace, and Cloud gives Google distribution ChatGPT can't match through apps alone. Enterprise adoption shows in the numbers, with 70% of Google Cloud users now deploying Gemini features.
The API developer count of 2.4 million active users, up 118% year-over-year, suggests growing production implementations beyond experimentation. That's the real test: are CTOs shipping features or running pilots?
For teams evaluating LLM providers, the token-per-minute metric provides a benchmark. Calculate your peak loads, compare batch versus streaming costs, and factor in the operational overhead of managing retries and rate limits. Google's enterprise tooling around Workspace and Cloud may justify higher per-token costs if it reduces integration work.
The fine print matters here: self-reported user figures vary across sources, with some third-party estimates putting Gemini MAUs lower. What's verifiable is the API growth and token processing capacity, which track with Google Cloud's 48% revenue increase. The infrastructure spending to support that throughput, not disclosed separately, will show up in the capital expenditure line items going forward.