LLM WORLD

Gemini 3 Pro

Google DeepMind · · 96 days after Imagen 4

Verified

2 of 2 published values traced to a source3 primary sources, 3 archived1 not disclosed by the lab1 not researched yet

Why this status

Every value this record asserts — date (Nov 18, 2025) and context window (1 million) — was found in an archived primary source, checked 2026-08-10. Parameter count is not recorded, so nothing is claimed about it. Modalities read from the launch announcement, which states: “built from the start to seamlessly synthesize information about any topic across multiple modalities, including text, images, video, audio and code.” Output is recorded as text: no image, audio or video generation is claimed anywhere in the primary sources, which labs advertise prominently when present. Capabilities evidenced from the primary sources: coding: “Gemini 3 Pro outperforms previous models in reasoning, multimodality, and coding benchmarks.” | agentic: “Gemini 2 laid the foundation for agentic capabilities and pushed the frontiers on reasoning and thinking , helping with more complex tasks and ideas, leading to Gemini 2.”.

What these states mean · How records are judged · Has this record changed?

Google's natively multimodal reasoning flagship with DeepThink and 1M context.

visioncodingagenticreasoningfunction callingstructured outputmultimodallong-contextflagship

Details

ReleasedNovember 18, 2025
stated in blog.google

Gemini 3 Pro was released on November 18, 2025.

CompanyGoogle DeepMind
FamilyGemini
Lineageafter Gemini 2.5 Pro
EraReasoning models
TypeLanguage model · LLM
ModalitiesText + Image + Audio + Video → Text
CapabilitiesVision, Coding, Agentic, Reasoning, Function calling, Structured output
WeightsProprietary
LicenceNot applicable — proprietary
Context window1.05M tokens
stated in blog.google, storage.googleapis.com

Gemini 3 Pro supports a context window of 1,048,576 tokens.

ParametersNot disclosed

Benchmarks

Each row is a dated claim by a named party, not a property of the model. A later revision adds a row rather than replacing one. This project publishes no composite or overall score.

BenchmarkScoreReported byDate
SWE-bench Verified76.2 Vendor reportedNovember 18, 2025
Terminal-Bench54.2 Vendor reportedNovember 18, 2025
MMMU81 Vendor reportedNovember 18, 2025
GPQA Diamond91.9 Vendor reportedNovember 18, 2025
Humanity’s Last Exam37.5 Vendor reportedNovember 18, 2025

Current information, elsewhere

This record is a historical one: every figure above is traced to a dated primary source. For what is true today — live pricing, current rankings, availability — these are better maintained than anything here could be.

Sources

Record status: verified · confidence 90/100 · 3 of 3 primary

Every value this record asserts — date (Nov 18, 2025) and context window (1 million) — was found in an archived primary source, checked 2026-08-10. Parameter count is not recorded, so nothing is claimed about it. Modalities read from the launch announcement, which states: “built from the start to seamlessly synthesize information about any topic across multiple modalities, including text, images, video, audio and code.” Output is recorded as text: no image, audio or video generation is claimed anywhere in the primary sources, which labs advertise prominently when present. Capabilities evidenced from the primary sources: coding: “Gemini 3 Pro outperforms previous models in reasoning, multimodality, and coding benchmarks.” | agentic: “Gemini 2 laid the foundation for agentic capabilities and pushed the frontiers on reasoning and thinking , helping with more complex tasks and ideas, leading to Gemini 2.”.

Primary means published by the organisation that made the model. Verified records require at least one.

Gemini family

  1. Gemini 1.0 Ultra December 6, 2023
  2. Gemini 1.0 Pro December 6, 2023
  3. Gemini 1.0 Nano December 6, 2023
  4. Gemini 1.5 Pro February 15, 2024
  5. Gemini 2.0 Flash December 11, 2024
  6. Gemini 2.5 Pro March 25, 2025
  7. Gemini 3 Pro November 18, 2025
  8. Gemini 3.1 Flash-Lite May 7, 2026
  9. Gemini 3.5 Flash May 19, 2026
  10. Gemini 3.6 Flash July 21, 2026
  11. Gemini 3.5 Flash-Lite July 21, 2026
  12. Gemini 3.7 Flash August 13, 2026
  13. Gemini 3.8 Flash September 2, 2026
  14. Gemini 3.8 Flash Cyber September 2, 2026
  15. Gemini 4 Argon September 30, 2026

What changed from Gemini 2.5 Pro

Compared with the previous release in this series. A field appears only when both records state a value — where one does not, it is listed as uncomparable rather than dropped, because a gap in our research is not a finding about the model.

Not comparable: Parameters (not disclosed for both releases); Modalities (Gemini 3 Pro evidences fewer than Gemini 2.5 Pro, which means not evidenced rather than removed)

Compare them side by side →

Browse models

Compare Gemini 3 Pro with another model →
See Gemini 3 Pro on the interactive timeline →