Gemini 2.0 Flash

Claims and evidence

• Released 11 December 2024Experimental release (Gemini 2.0 Flash Experimental) in the Gemini API, Google AI Studio and Vertex AI, and in the Gemini app. ● Primary source[1]Gemini 2.0 Flash [5]December 11, 2024
• Deprecated 18 February 2026 ● Primary source[5]February 18, 2026
• Retired 1 June 2026 ● Primary source[5]June 1, 2026 [6]Gemini 2.0 models [7]Retired models: gemini-2.0-flash
• Default model in the Gemini app 30 January 2025 ● Primary source[8]2025.01.30
• Generally available (gemini-2.0-flash-001) 5 February 2025 ● Primary source[3] [5]February 5, 2025 [6]Gemini 2.0 models

• Status Retired ● Primary source[5]June 1, 2026 [6]Gemini 2.0 models [7]Retired models: gemini-2.0-flash
• Replaced by Gemini 3.6 FlashRecommended replacement on the Gemini API deprecations page when checked on 2026-10-01. The Agent Platform page recommends gemini-3.1-flash-lite, and the June 1, 2026 release note pointed to gemini-3.5-flash or gemini-3.1-flash-lite. ● Primary source[6]Gemini 2.0 models

• Successor of Gemini-1.5-Flash-002Google says 2.0 Flash builds on and improves upon Gemini 1.5 Flash without naming a version; gemini-1.5-flash-002 was the current 1.5 Flash version at the time (the gemini-1.5-flash id pointed to it from 2024-11-21). ● Primary source[1]Gemini 2.0 Flash [4]Model Information, Description [5]November 21, 2024

• Change · Modality Adds native image and audio output, first limited to early-access partners; at general availability output was text only.Compared with Gemini-1.5-Flash-002 ● Primary source[1]Gemini 2.0 Flash [2]Model features
• Change · Tool use Natively calls tools such as Google Search and code execution, alongside user-defined functions.Compared with Gemini-1.5-Flash-002 ● Primary source[1]Gemini 2.0 Flash [9]3. Native tool use
• Change · Other Comes with a Multimodal Live API for real-time audio and video streaming input.Compared with Gemini-1.5-Flash-002 ● Primary source[9]4. Multimodal Live API
• Change · Architecture The model card says it builds on the sparse mixture-of-experts design of Gemini 1.5, with a refined architecture and new optimization methods.Compared with Gemini-1.5-Flash-002 ● Primary source[4]

• Input text, image, audio, video ● Primary source[4]Model Information, Inputs [1]Gemini 2.0 Flash
• Output textText output at general availability. Native image output and text-to-speech audio output were offered to early-access partners during the experimental phase; the model card calls image output experimental. ● Primary source[4]Model Information, Outputs [2]Model features
• Feature Tool useNative use of Google Search and code execution. ● Primary source[1]Gemini 2.0 Flash [9]3. Native tool use
• Feature Function calling ● Primary source[9]3. Native tool use
• Feature Real timeMultimodal Live API with real-time audio and video streaming input. ● Primary source[9]4. Multimodal Live API [4]Intended Usage and Limitations
• Context window 1,048,576 tokensThe developer post gives 1 million tokens; the model card also gives an output limit of 8,192 tokens. ● Primary source[4]Model Information, Inputs [2]Model features
• Parameters Not disclosedNot disclosed by Google. No claim madeNo source
• Access API, consumer appGemini API through Google AI Studio and Vertex AI; Gemini app. ● Primary source[1]Gemini 2.0 Flash; Gemini 2.0 available in Gemini app [3]

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 →