←Back to NewsAI News/BenchmarksnewsBenchmarksLlmsGoogle's Gemini 3.7 Flash Hits 84.6% on AGI Benchmark at 2.4x Lower CostGoogle's mid-tier reasoning model nearly clears the 85% target on ARC-AGI-2 at a quarter per task, redrawing the cost-performance frontier.SourceARC PrizePublishedAug 20, 2026, 6:04 PMAuthorAlphaSignal NewsroomRead1 min readGoogle's mid-tier reasoning model nearly clears the 85% target on ARC-AGI-2 at a quarter per task, redrawing the cost-performance frontier.Reporting is indexed from AlphaSignal. Rights remain with the original publisher and cited sources.Read original report ↗Next readsEpoch AI · newsEpoch Audits 15 AI Benchmarks and Flags Nine as FlawedLiquid AI · newsLiquid AI's LFM2 Beats GPT-5 and Claude on Aging Research TasksAlphaSignal · paperMax Planck's talkie-1930-13b Shattered People's Nostalgia for a Moral Past
