Google released Gemini 2.5 Pro Experimental on March 25, 2025, claiming top position on the LMArena leaderboard "by a significant margin" and leading on math, science, and coding benchmarks including 63.8% on SWE-Bench Verified for agentic code tasks. The model combines an enhanced base architecture with "thinking" capabilities that allow extended reasoning before response generation. The model ships with a 1 million token context window (2 million coming soon) and will be available on Vertex AI in coming weeks, with production pricing to follow. For financial institutions, this marks a continued escalation in frontier model capability claims and availability in production environments.
326 other tracked developments touching the same claims.