Gemini 1.5 Pro

Claims and evidence

• Announced 15 February 2024Released the same day in a limited private preview in AI Studio and Vertex AI. ● Primary source[1]
• Released 23 May 2024General availability of gemini-1.5-pro-001 in the Gemini API.Other sources give: 24 May 2024 [4]Retired models: gemini-1.5-pro-001 Release date of gemini-1.5-pro-001 on Vertex AI (now Gemini Enterprise Agent Platform). ● Primary source[3]May 23, 2024
• Retired 29 September 2025Shutdown of the gemini-1.5-pro model code in the Gemini API, which ended Gemini 1.5 Pro in that API. The release notes do not give a separate shutdown date for gemini-1.5-pro-001 there.Other sources give: 24 May 2025 [4]Retired models: gemini-1.5-pro-001 Retirement date of gemini-1.5-pro-001 on Vertex AI. ● Primary source[3]September 29, 2025
• Public preview in the Gemini API (gemini-1.5-pro-latest) in 180+ countries 9 April 2024 ● Primary source[3]April 9, 2024 [2]
• 2M-token context window generally available in the Gemini API 27 June 2024 ● Primary source[3]June 27, 2024

• Status Retired ● Primary source[3]September 29, 2025 [4]Retired models: gemini-1.5-pro-001
• Replaced by Gemini 2.5 FlashRecommended upgrade for gemini-1.5-pro-001 on Agent Platform as listed when checked on 2026-10-01. ● Primary source[4]Retired models: gemini-1.5-pro-001

• Successor of Gemini 1.0 Pro EditorialEditorial link along the Pro tier: Google presents Gemini 1.5 as the next generation after Gemini 1.0 and compares 1.5 Pro with 1.0 Pro and 1.0 Ultra, without naming a direct predecessor. ● Primary source[1]

• Change · Architecture Uses a Mixture-of-Experts architecture, which Google says makes the model more efficient to train and serve.Compared with Gemini 1.0 Pro ● Primary source[1]
• Change · Context length Standard context window grows from 32,000 to 128,000 tokens, with up to 1 million in private preview and 2 million generally available in the Gemini API from June 2024.Compared with Gemini 1.0 Pro ● Primary source[1] [3]June 27, 2024
• Change · Modality Accepts audio (speech) and video alongside text and images; at API launch Gemini 1.0 Pro took text only, with images through the separate Pro Vision endpoint.Compared with Gemini 1.0 Pro ● Primary source[2] [5]Gemini Pro is available today
• Change · Efficiency Google says 1.5 Pro reaches quality comparable to 1.0 Ultra while using less compute.Compared with Gemini 1.0 Pro ● Primary source[1]

• Input text, image, audio, videoAudio (speech) input reached the Gemini API with the public preview of 2024-04-09. ● Primary source[1]Greater context, more helpful capabilities [2]Unlock new use cases with audio and video modalities [6]Abstract
• Output text ● Primary source[7]1,000,000 tokens: Unlocking new use cases for developers
• Feature Long context ● Primary source[1]Greater context, more helpful capabilities
• Context window 128,000 tokensStandard window at announcement; a limited group could use up to 1 million tokens in the private preview. The Gemini API release notes add a 2M window (waitlist 2024-05-14, generally available 2024-06-27). ● Primary source[1]Introducing Gemini 1.5 [7]1,000,000 tokens: Unlocking new use cases for developers
• Parameters Not disclosedNot disclosed by Google. No claim madeNo source
• Access APIGemini API (Google AI Studio) and Vertex AI; private preview from 2024-02-15, public preview from 2024-04-09. ● Primary source[1]Build and experiment with Gemini models [8] [2]

Lineage

Variants

No variants recorded in this record.

Related AI Radar coverage

AI Radar coverage starts in June 2026; no coverage linked yet.

All model releases from Google on AI Radar →