Anthropic holds the strongest position among frontier labs heading into late June, driven by the recent release of Claude Opus 4.7 and its consistent edge on agentic and coding benchmarks. Traders see this as extending the company’s recent momentum in high-stakes reasoning tasks, where the model outperforms peers on expert-level workflows and self-improvement features introduced in early May. Google remains the clearest challenger with Gemini 3.1 Pro leading several independent reasoning and accuracy leaderboards, though its gains have not yet displaced Anthropic’s lead in the specific arena format traders track. OpenAI’s GPT-5.5 update improved factuality but has not produced the same benchmark surge, leaving it a distant third in current sentiment. With resolution tied to live leaderboard standings at month-end, any new model drop or major benchmark shift before June 30 could still alter the ordering.