Gemini 3.1 Pro Preview vs GPT-4O
As of Oct 14, 2026, Gemini 3.1 Pro Preview leads GPT-4O in 3 of 3 shared categories: Math, Science, and Abstraction (ARC).
Gemini 3.1 Pro Preview
Yes. Yes on Math (best per dollar). Not the overall leader.
GPT-4O
It used to be. Was SOTA in General LLM.
| Category | Gemini 3.1 Pro Preview | GPT-4O |
|---|---|---|
| General LLM | #15 · Was SOTA · General LLM Margin 10.204081632653057 LMArena leaderboard dataset · Oct 13, 2026 | #— · Was SOTA · General LLM Margin — LMArena leaderboard dataset · Feb 21, 2025 |
| Reasoning | #14 · Was SOTA · Reasoning Margin 10.273972602739732 LMArena leaderboard dataset · Oct 13, 2026 | #— · Not enough data Margin — LMArena leaderboard dataset · Feb 6, 2025 |
| Math | #3 · Near-SOTA · Math Margin 4.444444444444443 Epoch AI benchmarking hub · Feb 19, 2026 | #45 · #45 in Math Margin 100 Epoch AI benchmarking hub · Nov 20, 2024 |
| Science | #7 · #7 in Science Margin — Epoch AI benchmarking hub · Feb 19, 2026 | #131 · #131 in Science Margin — Epoch AI benchmarking hub · Nov 20, 2024 |
| Coding | #— · Not enough data Margin — LMArena leaderboard dataset · Oct 13, 2026 | #— · Not enough data Margin — LMArena leaderboard dataset · Feb 17, 2025 |
| Coding › Web dev | #58 · #58 in Coding › Web dev Margin 53.271028037383175 LMArena leaderboard dataset · Oct 13, 2026 | — |
| Long context | #12 · Was SOTA · Long context Margin 8.163265306122454 LMArena leaderboard dataset · Oct 13, 2026 | #— · Not enough data Margin — LMArena leaderboard dataset · Jan 22, 2025 |
| Vision | #25 · Was SOTA · Vision Margin 68.42105263157896 Epoch AI benchmarking hub · Feb 19, 2026 | #— · Not enough data Margin — LMArena leaderboard dataset · Oct 9, 2026 |
| Factuality | #18 · #18 in Factuality Margin 14.179104477611943 LMArena leaderboard dataset · Oct 13, 2026 | #— · Not enough data Margin — Epoch AI benchmarking hub · Aug 6, 2024 |
| Abstraction (ARC) | #2 · #2 in Abstraction (ARC) Margin — Epoch AI benchmarking hub · Feb 19, 2026 | #87 · #87 in Abstraction (ARC) Margin — Epoch AI benchmarking hub · Nov 20, 2024 |
Gemini 3.1 Pro Preview: $4.5 per 1M blended tokens.
GPT-4O: $4.375000000000001 per 1M blended tokens.
Gemini 3.1 Pro Preview: closed weights.
GPT-4O: closed weights.
Math: As of Oct 14, 2026, Gemini 3.1 Pro Preview leads GPT-4O in Math at #3 against #45.
Science: As of Oct 14, 2026, Gemini 3.1 Pro Preview leads GPT-4O in Science at #7 against #131.
Abstraction (ARC): As of Oct 14, 2026, Gemini 3.1 Pro Preview leads GPT-4O in Abstraction (ARC) at #2 against #87.
Gemini 3.1 Pro Preview · GPT-4O
Gemini 3.1 Pro Preview or GPT-4O?
As of Oct 14, 2026, Gemini 3.1 Pro Preview leads GPT-4O in 3 of 3 shared categories: Math, Science, and Abstraction (ARC).
Gemini 3.1 Pro Preview or GPT-4O for Math?
As of Oct 14, 2026, Gemini 3.1 Pro Preview leads GPT-4O in Math at #3 against #45.
Gemini 3.1 Pro Preview or GPT-4O for Science?
As of Oct 14, 2026, Gemini 3.1 Pro Preview leads GPT-4O in Science at #7 against #131.
Gemini 3.1 Pro Preview or GPT-4O for Abstraction (ARC)?
As of Oct 14, 2026, Gemini 3.1 Pro Preview leads GPT-4O in Abstraction (ARC) at #2 against #87.