Gemini 3 / 2.5 family · Nano Banana image · Veo video · Lyria music · agents & robotics models — Sources: Google AI for Developers official docs Model list · Pricing · Deprecations (synced 2026-08-30)
Direct recommendation by task type
Understand these patterns and instantly grasp any model name
| Element | Meaning | Example |
|---|---|---|
| gemini-X.Y version number | Higher number = newer and stronger | gemini-2.0 → 2.5 → 3 → 3.1 → 3.7 |
| -pro | Flagship enhanced, top reasoning | gemini-3.1-pro-preview, gemini-2.5-pro |
| -flash | Speed-optimized, balanced | gemini-3.7-flash, gemini-2.5-flash |
| -flash-lite | Cheapest and fastest lite | gemini-3.5-flash-lite, gemini-3.1-flash-lite |
| -preview | Preview (production-ready; 2-week deprecation notice) | gemini-3.1-pro-preview |
| -latest | Latest alias, hot-swaps with versions | gemini-flash-latest |
| nano-banana (🍌) | Image-generation family codename | gemini-3-pro-image (Pro), gemini-3.1-flash-image (2) |
| -image | Supports image output | gemini-3.1-flash-image |
| -live | Live API realtime audio/video chat | gemini-3.1-flash-live-preview |
| -native-audio | Native audio processing (Live API) | gemini-2.5-flash-native-audio-preview-12-2025 |
| -live-translate | Realtime speech translation | gemini-3.5-live-translate-preview |
| -tts | Text to speech | gemini-3.1-flash-tts-preview |
| -transcribe | Speech to text | gemini-3.5-transcribe(-live) |
| omni- | All-in-one video: generate/edit/interpolate/extend + native audio | gemini-omni-1.1-flash |
| veo-X | Video generation family | veo-3.1-generate-preview |
| lyria-X | Music generation family | lyria-3-pro-preview, lyria-3-clip-preview |
| robotics-er | Robotics embodied reasoning | gemini-robotics-er-2-preview |
| deep-research | Deep research agent | deep-research-max-preview-04-2026 |
| antigravity | Hosted general agent (autonomous in sandbox) | antigravity-preview-05-2026 |
| -computer-use | Computer use | gemini-2.5-computer-use-preview-10-2025 |
| gemma-X | Open-source lightweight family | gemma-4, gemma-3n |
| -001 / -exp | Stable version number / experimental | gemini-embedding-001, lyria-realtime-exp |
Latest generation; all support thinking (output price includes thinking tokens) and 1M context; cached input saves up to 90%, Batch at half price; listed prices are 2026 promotional rates (roughly 2x from 2027-01-01); all support PDF input
| Model ID | Tier | Price | Modality | Reasoning | Speed | Context | Input | Output | Notes |
|---|---|---|---|---|---|---|---|---|---|
| gemini-3.1-pro-previewPreview | Flagship | Best | Standard | 1M | 1M | 64K | Current top reasoning, agentic & complex coding, $2/$12 | ||
| gemini-3.7-flashNEWRec | Flagship | Deep | Fast | 1M | 1M | 64K | Newest, strongest Flash (Stable), complex coding & multi-step execution, $0.75/$3.75 | ||
| gemini-3.6-flash | Balanced | Deep | Fast | 1M | 1M | 64K | Previous Flash (Stable), enhanced search & grounding, $0.75/$3.75 | ||
| gemini-3.5-flash | Balanced | Standard | Fast | 1M | 1M | 64K | Base Flash (Stable), high-throughput general tasks, $1.5/$9 | ||
| gemini-3.5-flash-lite | Budget | Basic | Fastest | 1M | 1M | 64K | Fastest, cheapest 3.5 (Stable), $0.3/$2.5 | ||
| gemini-3.1-flash-lite | Budget | Basic | Fastest | 1M | 1M | 64K | Frontier-grade at budget price (Stable), $0.25/$1.5; announced retirement 2027-05 | ||
| gemini-3-flash-preview | Balanced | Deep | Fast | 1M | 1M | 64K | Early preview, $0.5/$3; migrating to gemini-3.6-flash recommended | ||
| gemini-3-pro-previewDeprecated | Flagship | Best | Standard | 1M | 1M | 64K | Gemini 3 Pro preview, shut down (2026-03-09); migrate to gemini-3.1-pro-preview | ||
| gemini-3.1-flash-lite-previewDeprecated | Budget | Basic | Fastest | 1M | 1M | 64K | 3.1 Lite preview, shut down (2026-05-25); migrate to gemini-3.1-flash-lite |
Current production workhorses (Stable), 1M context, thinking & Function Calling; all support PDF input
| Model ID | Tier | Price | Modality | Reasoning | Speed | Context | Input | Output | Notes |
|---|---|---|---|---|---|---|---|---|---|
| gemini-2.5-proRec | Flagship | Best | Standard | 1M | 1M | 64K | Strongest of the 2.5 generation, top choice for complex tasks & coding, $1.25/$10 | ||
| gemini-2.5-flashRec | Balanced | Standard | Fast | 1M | 1M | 64K | Best value, low latency high throughput, $0.3/$2.5 | ||
| gemini-2.5-flash-lite | Budget | Basic | Fastest | 1M | 1M | 64K | Cheapest of the 2.5 generation, large-scale calls, $0.1/$0.4 | ||
| gemini-2.5-flash-image | Balanced | Standard | Fast | 64K | 64K | 32K | Previous Nano Banana (🍌) image generation, $0.3 input; announced retirement 2026-10 | ||
| gemini-2.5-computer-use-preview-10-2025Preview | Balanced | Deep | Slow | 128K | 128K | 64K | Dedicated computer use: watch screen, click/type/navigate, $1.25/$10 |
Speech-to-text, text-to-speech, realtime chat & realtime translation; Live series streams bidirectionally over WebSocket
| Model ID | Tier | Price | Type | Input | Output | Speed | Features | Notes |
|---|---|---|---|---|---|---|---|---|
| gemini-3.5-transcribeNEWRec | Flagship | Speech to text | Fast | Speaker diarization / word-level timestamps | Low-latency transcription, $2/$12 | |||
| gemini-3.5-transcribe-liveNEW | Flagship | Live transcription | Fastest | Streaming live | Realtime streaming transcription, $3.5/$21 | |||
| gemini-3.1-flash-tts-previewNEW | Balanced | Text to speech | Fast | Fine-grained audio tags control | Latest TTS, $1/$20 | |||
| gemini-2.5-flash-preview-ttsPreview | Balanced | Text to speech | Fastest | Style/speed controllable | Previous TTS, $0.5/$10 | |||
| gemini-2.5-pro-preview-ttsPreview | Flagship | Text to speech | Fast | High fidelity | Podcast/audiobook grade, $1/$20 | |||
| gemini-3.1-flash-live-previewNEW | Balanced | Realtime chat | Fastest | Live API | Realtime voice chat flagship, $0.75/$4.5 | |||
| gemini-2.5-flash-native-audio-preview-12-2025Preview | Balanced | Realtime chat | Fastest | Native audio reasoning | Previous Live API model, $0.5/$2 | |||
| gemini-3.5-live-translate-previewNEW | Flagship | Realtime translation | Fastest | 70+ languages | Speech-to-speech realtime translation, $3.5/$21 |
Nano Banana is the current image-generation workhorse; Imagen 4 has been merged and shut down
| Model ID | Tier | Price | Input | Output | Speed | Resolution | Notes |
|---|---|---|---|---|---|---|---|
| gemini-3-pro-imageRec | Flagship | Standard | up to 4K | Nano Banana Pro: SOTA image generation & editing, complex layouts/text rendering, $2/$12 | |||
| gemini-3.1-flash-image | Balanced | Fast | 2K | Nano Banana 2: high-efficiency volume generation, $0.5/$3 | |||
| gemini-3.1-flash-lite-image | Budget | Fastest | Standard | Nano Banana 2 Lite: ultra-low latency & cost, $0.25/$1.5 | |||
| gemini-2.5-flash-image | Balanced | Fast | Standard | Previous Nano Banana (🍌), $0.3 input; announced retirement 2026-10 | |||
| imagen-4.0-generate-001Deprecated | Balanced | Standard | 2K | Shut down 2026-08-17; migrate to Nano Banana |
Billed by output video seconds, includes synced audio; Veo 3.1 supports up to 4K, Omni supports generate/edit/interpolate/extend
| Model ID | Tier | Type | Billing | Speed | Resolution | Notes |
|---|---|---|---|---|---|---|
| veo-3.1-generate-previewPreview | Flagship | Text to video | $0.4-0.6/sec | Slowest | 1080p-4K | Cinematic quality + native synced audio |
| veo-3.1-fast-generate-previewRec | Balanced | Text to video | $0.1-0.3/sec | Standard | 1080p-4K | Fast generation, best value |
| veo-3.1-lite-generate-previewNEW | Budget | Text to video | $0.05-0.08/sec | Fast | 1080p | Developer-friendly low-cost version (no 4K) |
| gemini-omni-1.1-flashNEW | Balanced | Video generate/edit/interpolate/extend | $1.5/$9 (token) | Standard | — | Native audio, most complete features |
| veo-3.0-generate-001Deprecated | Balanced | Text to video | $0.2-0.4/sec | Slowest | 1080p | Shut down 2026-06-30; migrate to Veo 3.1 |
| gemini-omni-flash-previewDeprecated | Budget | Video generate/edit | $1.5/$9 (token) | Standard | — | Preview; retires 2026-09-30; migrate to gemini-omni-1.1-flash |
Billed per generated song
| Model ID | Tier | Type | Billing | Duration | Notes |
|---|---|---|---|---|---|
| lyria-3-pro-previewNEWRec | Flagship | Full songs | $0.08/song | Full track | Flagship music generation, complex structure integrity |
| lyria-3-clip-previewNEW | Balanced | Clips / loops | $0.04/song | 30 sec | Clips, loops and previews |
| lyria-realtime-exp | Balanced | Realtime streaming | Experimental | Streaming | High-fidelity realtime music generation, fine-grained creative control |
Agent models with autonomous planning & execution; preview stage, pricing TBD by Google
| Model ID | Tier | Price | Use | Notes |
|---|---|---|---|---|
| deep-research-max-preview-04-2026Preview | Flagship | — | Deep research | Strongest automated research: retrieves & synthesizes hundreds of sources, produces interactive cited reports |
| deep-research-preview-04-2026Preview | Balanced | — | Deep research | Autonomously plans and executes multi-step research |
| antigravity-preview-05-2026Preview | Flagship | — | General agent | Autonomous planning/reasoning/code execution/file management/web browsing in an isolated Linux sandbox |
Embodied reasoning: understand physical space, plan multi-step tasks, multi-robot collaboration
| Model ID | Tier | Price | Modality | Reasoning | Speed | Context | Input | Output | Notes |
|---|---|---|---|---|---|---|---|---|---|
| gemini-robotics-er-2-previewNEWRec | Flagship | Deep | Fast | 128K | 128K | 64K | Latest embodied reasoning: video understanding/spatial reasoning/multi-step tool orchestration/multi-robot collaboration, $2/$10 | ||
| gemini-robotics-er-1.6-previewDeprecated | Balanced | Deep | Fast | 128K | 128K | 64K | Previous ER, retires 2026-08-31, migrate to ER 2, $1/$5 | ||
| gemini-robotics-er-1.5-previewDeprecated | Balanced | Deep | Fast | 128K | 128K | 64K | Previous ER, shut down (2026-04-30); migrate to gemini-robotics-er-1.6-preview |
Semantic search & RAG; Embedding 2 is the first multimodal embedding model (unified vector space for text/image/audio/video/PDF)
| Model ID | Tier | Price | Type | Input | Speed | Features | Notes |
|---|---|---|---|---|---|---|---|
| gemini-embedding-2NEWRec | Flagship | Multimodal embedding | Standard | Text $0.2 / Image $0.45 / Audio $6.5 / Video $12 | First multimodal Embedding, unified vector space | ||
| gemini-embedding-001 | Balanced | Embed | Standard | $0.15/1M | High-dim text vectors, retires 2028-05-14 |
Lightweight open-source models sharing Gemini's lineage, free to use, self-hostable
| Model ID | Tier | Price | Modality | Reasoning | Speed | Context | Notes |
|---|---|---|---|---|---|---|---|
| gemma-4NEWOSS | Budget | Open source & free | Standard | Fastest | 128K | Latest open-source flagship, same lineage as Gemini | |
| gemma-3OSS | Budget | Open source & free | Standard | Fastest | 128K | Text/image understanding & reasoning | |
| gemma-3nOSS | Budget | Open source & free | Basic | Fastest | 32K | Multimodal input, optimized for mobile | |
| gemma-2OSS | Budget | Open source & free | Standard | Fast | 8K | Text generation/summarization/extraction (9B/27B) | |
| codegemmaOSS | Budget | Open source & free | Standard | Fast | 8K | Code generation & completion (2B/7B) | |
| ShieldGemmaOSS | Budget | Open source & free | Standard | Fast | — | Safety classification & content filtering | |
| TxGemmaOSS | Budget | Open source & free | Standard | Fast | — | Therapy prediction/drug discovery/genomics | |
| MedGemmaOSS | Budget | Open source & free | Standard | Fast | — | Medical imaging analysis (DICOM) | |
| paligemmaOSS | Budget | Open source & free | Standard | Fast | — | Image captioning & visual QA (3B) |
All shut down 2026-06-01; migrate to the corresponding Gemini 3.x replacements; all support PDF input
| Model ID | Tier | Price | Modality | Reasoning | Speed | Context | Input | Output | Notes |
|---|---|---|---|---|---|---|---|---|---|
| gemini-2.0-flashDeprecated | Balanced | Standard | Fast | 1M | 1M | 8K | Shut down (2026-06-01); migrate to gemini-3.6-flash | ||
| gemini-2.0-flash-liteDeprecated | Budget | Basic | Fastest | 1M | 1M | 8K | Shut down (2026-06-01); migrate to gemini-3.1-flash-lite |
Retirement info for the models covered on this page (shutdown dates are the earliest official dates). Source: Gemini official deprecation schedule (synced 2026-08-30)
| Model ID | Lifecycle | Shutdown date | Replacement | Migration advice |
|---|---|---|---|---|
| gemini-3-pro-previewDeprecated | Retired | 2026-03-09 | →gemini-3.1-pro-preview | Shut down; drop-in replacement |
| gemini-robotics-er-1.5-previewDeprecated | Retired | 2026-04-30 | →gemini-robotics-er-1.6-preview | Shut down; drop-in replacement |
| gemini-3.1-flash-lite-previewDeprecated | Retired | 2026-05-25 | →gemini-3.1-flash-lite | Shut down; drop-in replacement |
| gemini-2.0-flashDeprecated | Retired | 2026-06-01 | →gemini-3.6-flash | Shut down; drop-in replacement |
| gemini-2.0-flash-liteDeprecated | Retired | 2026-06-01 | →gemini-3.1-flash-lite | Shut down; drop-in replacement |
| imagen-4.0-generate-001Deprecated | Retired | 2026-08-17 | →gemini-3.1-flash-image | Shut down; migrate to Nano Banana 2 |
| veo-3.0-generate-001Deprecated | Retired | 2026-06-30 | →veo-3.1-generate-preview | Shut down; drop-in replacement |
| gemini-robotics-er-1.6-previewDeprecated | Deprecated | 2026-08-31 | →gemini-robotics-er-2-preview | Retiring soon; migrate promptly |
| gemini-omni-flash-previewDeprecated | Deprecated | 2026-09-30 | →gemini-omni-1.1-flash | Preview; retires on schedule |
| gemini-2.5-flash-image | Deprecated | 2026-10-02 | →gemini-3.1-flash-image | Retirement announced; plan migration |
| gemini-3.1-flash-lite | Deprecated | 2027-05-07 | →gemini-3.5-flash-lite | Stable but announced retirement; plan migration |
| gemini-embedding-001 | Deprecated | 2028-05-14 | →gemini-embedding-2 | Still usable (until 2028); Embedding 2 recommended for new projects |
Match the best model to your needs quickly
| Scenario | Recommended | Alternatives | Key capability |
|---|---|---|---|
| Chat / Writing | gemini-3.7-flash | gemini-3.6-flash | 1M context $0.75/$3.75 |
| Complex reasoning / Math | gemini-3.1-pro-preview | gemini-2.5-pro | Top |
| Agentic coding workflows | gemini-3.7-flash | gemini-2.5-pro | + multi-step execution |
| Ultra-low cost | gemini-3.1-flash-lite | gemini-2.5-flash-lite | $0.25/$1.5 |
| Long-document processing~1M tokens | gemini-2.5-pro | gemini-3.7-flash | 1M context |
| Multimodal understanding (image+audio+video) | gemini-3.7-flash | gemini-2.5-flash | |
| Realtime voice chat | gemini-3.1-flash-live-preview | gemini-2.5-flash-native-audio-preview-12-2025 | Live API two-way streaming |
| Speech to text | gemini-3.5-transcribe | gemini-3.5-transcribe-live | Speaker diarization / word-level timestamps |
| Text to speech | gemini-3.1-flash-tts-preview | gemini-2.5-pro-preview-tts | Fine-grained audio tags control |
| Realtime speech translation | gemini-3.5-live-translate-preview | — | 70+ languages |
| High-quality image generation | gemini-3-pro-image | gemini-3.1-flash-image | up to 4K · Nano Banana Pro |
| Video generation | veo-3.1-fast-generate-preview | veo-3.1-generate-preview | 4K + synced audio |
| Music creation | lyria-3-pro-preview | lyria-3-clip-preview | $0.04-0.08/song |
| Semantic search / RAG | gemini-embedding-2 | gemini-embedding-001 | Multimodal unified vector space |
| Computer use automation | gemini-2.5-computer-use-preview-10-2025 | — | Watch screen, perform UI actions |
| Robotics control | gemini-robotics-er-2-preview | gemini-robotics-er-1.6-preview | Embodied reasoning · multi-robot collaboration |
| Private / edge deployment | gemma-4 | gemma-3n | Open source & free |