Grok is now the most popular model on OpenRouter (openrouter.ai)

OpenRouter’s usage dashboards indicate that Grok has climbed to the top spot as the most popular model on the platform, leading in token consumption and dominating model-share metrics by author and by primary use case. The shift shows up across multiple telemetry slices — overall token usage, per-author share, tool integrations, and even image-processing counts — and is reinforced by several large public apps opting into OpenRouter’s usage tracking. In short, Grok isn’t just getting more calls; it’s being chosen for a wider set of workloads and workflows than competing models. For the AI/ML community this matters because platform-level adoption shapes downstream tooling, cost dynamics, and where engineers optimize infrastructure and research effort. Higher token share implies real-world preference for Grok’s tradeoffs (latency, cost, instruction-following or multimodal capabilities) and drives ecosystem effects: more adapters, wrappers, benchmarks, and safety evaluations will center on it. The metrics around tool usage and images processed also highlight growing multimodal and tool-enabled patterns (retrieval, function calling, chaining) that practitioners should account for when designing pipelines, estimating inference cost, or prioritizing robustness and alignment testing.