←Back to NewsAI News/LlmsmodelLlmsAgentsTokenRhythm's NeoHorse-1 Beats Its Base Model by 6 Points on Agentic BenchmarksTokenRhythm fine-tunes Qwen3.5-4B with a routing harness that turns agent traces into training data, gaining 5.93 points across ten benchmarks.SourceAlphaSignalPublishedSep 5, 2026, 8:21 AMAuthorAlphaSignal NewsroomRead1 min readTokenRhythm fine-tunes Qwen3.5-4B with a routing harness that turns agent traces into training data, gaining 5.93 points across ten benchmarks.Reporting is indexed from AlphaSignal. Rights remain with the original publisher and cited sources.Read original report ↗Next readsOpenAI · newsOpenAI's GPT-6 Astra Hits 99.9% on AGI Benchmarks and Runs Your ComputerAdham Khaled · deep-diveCoding Agent Cost: Why Sessions Get ExpensiveDeepLearning.AI · newsXiaohongshu's Self-GC Cuts Agent Token Costs by 20% Without Losing Critical Context
