Gemini 3 Pro
Gemini 3 Pro is a closed AI model from Google with an index score of 160 (GPT-4 = 100), rank 53 of 217.
160145–176Index score#53 of 217 models
–Price per million tokensno public API price found
–Context windowtokens it reads at once
18 Nov 2025ReleasedReplaced by Gemini 3.1 Pro
Where it sitson the frontier
Among the first models to reach this score.
Test resultsby area · tick = best by any model
Reasoning & knowledgeevidence
Mathsevidence
Codingevidence
Agentsevidence
Professional workevidence
Novel problemsevidence
Long tasksevidence
Third-party results only, at the model’s best standard setting. Hover a bar for who ran it.
What others sayrank on 12 leaderboards
The Model Indexthird-party tests, our method#53of 217 · 160
LMArena · Creative writingcreative writing#12of 360 · 1484LMArena · Textoverall chat quality, judged by people#19of 362 · 1486LMArena · Long promptslong, complex requests#26of 340 · 1491LMArena · Visionunderstanding images#12of 145 · 1289LMArena · Instruction followingfollowing detailed instructions#32of 362 · 1473LMArena · Codingcoding questions in chat#41of 357 · 1517LMArena · Mathsmaths questions in chat#40of 346 · 1475Epoch Capabilities Indexoverall capability from many benchmarks#55of 231 · 153.0LMArena · Searchanswering with web search#17of 41 · 1207LMArena · WebDevbuilding web apps and front ends#74of 133 · 1441LMArena · Documentsworking with long documents#31of 40 · 1434Best attop three among current models
Where to use itprices checked 18 Nov 2026
Host prices from OpenRouter; going direct can differ.
Facts and sources
Released18 Nov 2025
Announcementblog.google ↗
Release entryGemini 3 Pro
Scored asGemini 3 Pro
Output speed–
Hugging Faceclosed weights