Marina Lyubimova
⬤ Kilo Code shot back to first place on the OpenRouter daily rankings after an intense period of activity drove the platform to process roughly 313 billion tokens within 24 hours. The platform overtook OpenClaw and reasserted its position as the leading AI coding tool by overall usage volume.
⬤ The lion's share of that traffic came from GLM-5, Z.ai's latest model, which alone accounted for approximately 222 billion tokens during the same window. Screenshots from the leaderboard confirmed Kilo Code's dominance, with OpenClaw trailing in second place. The spike reflects surging demand across coding agents and developer-focused platforms, particularly as discussions around GLM-5 coding benchmark results gained traction in developer communities.
⬤ Z.ai offered temporary free access to GLM-5 through Sunday, which coincided directly with the massive jump in token processing. The announcement claimed the model outperformed Opus 4.5 and 4.6 across multiple benchmarks, sparking interest among developers eager to test its capabilities.
⬤ The rankings shift also highlighted ongoing competition in the AI coding space, where platforms constantly jockey for position. Recent Grok code usage ranking comparisons showed similar volatility, underscoring how quickly user preferences and workload patterns can pivot based on model releases and promotional access windows.
⬤ These rapid swings in token volumes signal expanding demand for inference workloads across the large language model ecosystem. As compute infrastructure providers race to support heavier traffic, the competition between AI development platforms continues to intensify, with each new model release potentially reshuffling the leaderboard overnight.
Marina Lyubimova
