Google DeepMind·Gemini

Gemini 2.5 Pro

Google built “thinking” directly into the general Gemini generation, with emphasis on complex reasoning and coding.

LegacyLanguageReasoningCodingAgentMultimodalWeights not released

MODEL NUMBER019

First public date: 2025-03-25
Organization: Google DeepMind
Family: Gemini

Specifications

Only published specifications are shown.

Context: 1,000,000 tokens
Maximum input: 1,048,576 tokens↗
Maximum output: 65,536 tokens↗
Input modalities: Text · Image · Audio · Video
Open weights: No
Status: Legacy

Overview

Release dates

March 25, 2025 — Gemini 2.5 Pro Experimental launched with thinking built into the new generation. ↗ June 17, 2025 — Gemini 2.5 Pro became generally available. ↗

Why it mattered

Gemini 2.5 moved visible reasoning from a Flash experiment into the default direction for subsequent Gemini models.

Specifications alongside predecessor

Compared with Gemini 1.5 Pro ↗

PreviousGemini 1.5 Pro— · 2M ctx

CurrentGemini 2.5 Pro— · 1M ctx

Architecture & capabilities

Model capabilities

Thinking modeCodingAgentLong context

Release history

2025-03-25Preview

Gemini 2.5 Pro Experimental launched with thinking built into the new generation.

View source

2025-06-17General release

Gemini 2.5 Pro became generally available.

View source

Related models

Select a card to view the related model.

PredecessorGemini 1.5 ProView details →SuccessorGemini 3 Pro PreviewView details →SiblingGemini 2.5 FlashView details →

Sources

Gemini 2.5: Our most intelligent AI modelGoogle · Published 2025-03-25

Official · Verified 02

Gemini API model limitsGoogle · Checked 2026-09-07

Official · Verified 03

Gemini 2.5 model family expandsGoogle · Published 2025-06-17 · Checked 2026-09-03

Official · Verified

Suggest a correction

Copy this template, add the field, proposed value, and primary source, and share it with the archive maintainer.

Model ID: gemini-2-5-pro
Page: https://llmlineage.com/en/models/gemini-2-5-pro
Field:
Proposed value:
Primary source URL:
Reason: