Google#14
Overall rank
14.7
Podium score
1
Board wins
Feb 19, 2026
Released
Chat
Which model people prefer in blind head-to-head conversations.
Reasoning
Hard questions with checkable answers: science, math, puzzles, expert exams.
1
MMLU-Pro3
HLE5
GPQA Diamond13
ARC-AGI-222
LiveBench29Epoch ECI30
AA Intelligence
MMLU-Pro91.2% · of 100 · stale since 2026-03-11
as “Gemini-3.1-Pro”
HLE46.4% · of 42
as “gemini-3.1-pro-preview (thinking high)”
GPQA Diamond94.4% · of 100
as “gemini-3.1-pro-preview_high”
ARC-AGI-277.1% · of 78
as “Gemini 3.1 Pro (Preview)”
LiveBench77.0% · of 59
as “gemini-3.1-pro-preview-high”
Epoch ECI154.9 · of 100
as “Gemini 3.1 Pro”
AA Intelligence29.7 · of 100
as “Gemini 3.1 Pro Preview”
Coding
Fixing real repositories, building web apps, editing code.
Agents and tools
Calling functions and tools, using MCP servers, searching the web.
Vision
Understanding images, charts and documents.
More from Google
Gemini 3.8 FlashGemini 3.7 FlashGemini 3.6 FlashGemini 3.5 FlashGemini 3.1 Flash-LiteGemini 3Gemini 3 ProGemini 3.8 Flash TTS
Scores as published by each board on the capture date. Model names and logos belong to their owners; logos via logo.dev.