Compare / Gemini 2.5 Pro vs Gemini 3.1 Pro Preview

Gemini 2.5 Pro vs Gemini 3.1 Pro Preview

16 results from 4 sources that measured both models. Each row is in the source's own units; there is no overall winner.

Gemini 2.5 Pro
Google · profile
Gemini 3.1 Pro Preview
Google · profile
Results compared
16
Shared sources
4
BenchmarkMetricGemini 2.5 ProGemini 3.1 Pro Preview
Arena (formerly LMArena)reportedOverallArena rating 1,4311,459
Arena (formerly LMArena)reportedOverallArena rating 1,1421,208
Arena (formerly LMArena)reportedOverallArena rating 1,4621,472
Arena (formerly LMArena)reportedBusiness, management and financeArena rating 1,4371,477
Arena (formerly LMArena)reportedCreative writingArena rating 1,4431,480
Arena (formerly LMArena)reportedExpert promptsArena rating 1,4581,509
Arena (formerly LMArena)reportedInstruction followingArena rating 1,4381,479
Arena (formerly LMArena)reportedOverallArena rating 1,4461,487
Arena (formerly LMArena)reportedWriting, literature and languageArena rating 1,4451,480
tau2-benchreportedConsistency (pass^4)% of tasks 1.0%9.3%
tau2-benchreportedTask success (pass^1)% of tasks 13.7%26.0%
UGI LeaderboardreportedRequested-length error% off the requested word count 19.0%19.0%
UGI LeaderboardreportedStyle adherencescore from 0 to 1 0.400.34
UGI LeaderboardreportedWriting scorescore out of 100 65.272.2
Vectara Hallucination LeaderboardreportedAnswer rate% of documents 99.1%99.4%
Vectara Hallucination LeaderboardreportedHallucination rate% of summaries 7.0%10.4%

Bold green marks the better value on that metric. Where a source reports ranges that overlap, the difference may not be meaningful; see the benchmark page for ranges.