Google · proprietary

Gemini 3 Pro

Gemini 3 Pro by Google appears in 3 sources with Reasoning at 49.76. Best read for Code quality, Coding, High intelligence.

CreatorGoogle

Release date2025-11-18

Knowledge cutoffNot published

ContextUnknown

Input priceUnknown

Output priceUnknown

Modalitytext + vision

CountryUS

Metrics

All source-backed metrics

Open in leaderboard

LLM Stats Rank11ranking · source_rank

Reasoning49.76reasoning · index_reasoning

Math51.75math · index_math

Coding32.36coding · index_code

Vision34.08multimodal · index_vision

Tool calling18.97tool_calling · index_tool_calling

Long context8.82long_context · index_long_context

Healthcare49.45domain · index_healthcare

GPQA91.9 %reasoning · gpqa_score

AIME 2025100 %math · aime_2025_score

SWE-bench Verified76.2 %coding · swe_bench_verified_score

Code Arena1,579.31 %coding · coding_arena_score

Humanity Last Exam45.8 %reasoning · hle_score

ARC-AGI v231.1 %reasoning · arc_agi_v2_score

MMMLU91.8 %reasoning · mmmlu_score

MMMU-Pro81 %multimodal · mmmu_pro_score

SimpleQA72.1 %knowledge · simpleqa_score

MRCR v226.3 %long_context · mrcr_v2_score

CharXiv-R81.4 %multimodal · charxiv_r_score

ScreenSpot Pro72.7 %multimodal · screenspot_pro_score

Arena Rating1,479.66metric

Arena Rank5metric

Vote Count41,331metric

GPQA Diamond91.9 %metric

AIME 2025100 %metric

Humanity's Last Exam45.8 %metric

Evidence

Citations and source overlap

Arena AILM Arena text leaderboard (overall) mirrored in the public leaderboard dataset.Retrieved 2026-05-20 VellumTop models per benchmark from page-level structured data (JSON-LD).Retrieved 2026-05-20 LLM StatsPublic full LLM leaderboard advanced explorer snapshot, including ranking, benchmark indexes, pricing, speed, context, license, and model metadata.Retrieved 2026-06-01

FAQ

How should I read this profile?

Treat this as a source-backed model dossier, not an EvalKit-run verification. The public values are replicated from linked sources and kept source-scoped.

Is Gemini 3 Pro verified by EvalKit?

No. EvalKit currently shows 0 verified rows until real run evidence exists.

Why can metrics disagree?

Different sources test different tasks, dates, prompts, and aggregation methods. EvalKit keeps those differences visible instead of merging them into a fake universal score.