Google disclosed 18 benchmarks for Gemini 4 Argon, and VentureBeat counted it as the outright leader on 12 and tied for first on one, ahead of OpenAI's GPT-6 Astra and Anthropic's Claude Opus 5.5 in total top scores. It scored 19.6% on Harvey's Legal Agent Benchmark against 5.4% for GPT-6 Astra and 3.8% for Claude Opus 5.5. It trailed on Terminal-Bench 4.0, where Opus 5.5 scores 66.4% against Argon's 57.4%.
Artificial Analysis assigns Argon a score of 53 on its Intelligence Index. VentureBeat notes the comparisons are Google's previously disclosed figures rather than a new independently replicated set, so buyers should test on their own workloads.
Source: VentureBeat