● ModelsNov 2, 2026 · 13:518 sources
Google's Gemini 4 Carbon Tops Mercor's APEX-Agents Leaderboard With First Mean Score Above 90%
From marketbrief.now
AI;DR
Google's Gemini 4 Carbon debuted at No. 1 on Mercor's APEX-Agents leaderboard with a 90.4% mean. Carbon achieved the only evaluated model mean above 90% across multi-step tasks in investment banking, management consulting, and corporate law. Perplexity CEO Aravind Srinivas stated early evaluations showed Gemini 4 Carbon set a state-of-the-art score on the Wide-And-Deep-Research (WANDR) benchmark.
Read originalSign in to suggest
Sources
+ 2 more
- SUPPORT@tulseedoshi“We're starting to roll out Carbon to a subset of cyber defenders through our Fairwind Program, and we’ll continue to iterate before expanding access!”x.com
- SUPPORT@paradite_“gemini 4 carbon didn't beat claude opus 5.5 in terminal-bench 4.0. carbon scores 65.2%, below opus 5.5 (66.4%). but google let artificial analysis run tb4 benchmark, which put opus 5.5 below carbon by lowering opus 5.5 score.”x.com