Gemini 3 Flash Preview
gemini-3-flash-preview- 🆕 Google's fast, thinking-enabled Flash model, released December 2025.
- 🧠 Combines Gemini 3 Pro-style reasoning with Flash-line low latency.
- 📏 256,000-token context window per the catalog listing.
- 👁️ Multimodal inputs including vision and audio.
- 🔧 Function calling and tool use for agentic workflows.
- 🎯 Built for agentic workflows, multi-turn chat, and coding.
- ⚡ Positioned to deliver near-Pro reasoning with lower latency.
- 🌐 Includes web-search capability in its feature set.
Google is an American multinational technology corporation and one of the world's most valuable brands. A subsidiary of parent company Alphabet Inc., Google operates across search, cloud computing, consumer electronics, and artificial intelligence. Its DeepMind and Google…
Explore 17 more models by Google →Gemini 3 Flash Preview is Google's speed-and-value tier within the Gemini 3 family, engineered to pair thinking-style reasoning with the low latency and efficiency that define the Flash line. Google positions it as combining Gemini 3 Pro's reasoning capabilities with Flash-level latency, aiming at everyday tasks as well as more complex agentic workflows.
According to the catalog listing, the model handles multimodal inputs — spanning vision and audio alongside text — and exposes a 256,000-token context window. Its documented capabilities include reasoning, function calling, and web search, making it suited to long-running agent loops, multi-turn chat, and collaborative coding.
The model is offered as a preview under the model id gemini-3-flash-preview, released in December 2025. As a Flash-tier release, it sits below Google's Pro variants in the Gemini 3 lineup while emphasizing throughput and responsiveness for interactive development.
Within the broader lineup, it was later followed by the newer sibling [[sibling:gemini-3-5-flash|Gemini 3.5 Flash]], which continues the Flash series. For interactive development and multimodal understanding, Google describes the Flash tier as delivering strong reasoning and tool-use performance with lower latency than larger Gemini variants.
This About section is AI-generated from public sources via VeniceStats + Venice inference, with no human editing. It may contain inaccuracies.
| Seller | Reputation↓ | Routing | Input $/M | Cached $/M | Output $/M | Categories | API |
|---|---|---|---|---|---|---|---|
| Venice.ai Proxy 0x1f22…18c9 | 99 | #3 | $0.35 | $0.035 | $1.875 | chat,reasoning,audio,vision,video,multimodal,web-search | openai-chat-completions |
| ▲ Apex Ant 0x73b4…e736 | 83 | #2 | $0.315 | $0.0315 | $1.6875 | chat,fast,vision,multimodal,reasoning,long-context,agents | openai-chat-completions |
| surplusintelligence.ai 0x0e49…8927 | 68 | #1 | $0.2625 | $0.0263 | $1.4063 | agents,audio,chat,cheap,fast,frontier,function-calling,multimodal,reasoning,research,tasks,tools,video,vision,web-search | openai-chat-completions |
| NovaRoute AI 0xc50d…ed7b | 60 | gated | $0.2599 | $0.2599 | $1.3922 | chat,coding,code,reasoning,tasks,gemini,google,value,surplus,openai-compatible,low-cost,verified,github,response-auth,base-usdc,monitored | openai-chat-completions |
| Fire Ant 🔥🐜 0xbe05…bc5d | 37 | gated | $0.6248 | $0.061 | $3.3474 | base-usdc,chat,code,coding,gemini,github,google,json,low-cost,math,monitored,multimodal,openai-compatible,reasoning,response-auth,surplus,tasks,tools,value,verified | — |
| Meridian AI 0x8c8c…06f5 | 22 | gated | $0.1074 | $0.1074 | $0.6441 | chat,coding,multimodal | openai-chat-completions |
| antseed-neon-puma-944e 0x6650…944e | 3 | gated | $0.2387 | $0.07 | $1.2788 | chat,math | openai-chat-completions |
| antseed-opal-badger-2580 0xc85d…2580 | 1 | gated | $0.35 | $0.35 | $1.875 | chat | openai-chat-completions |
| AntFeed 0xddb6…1442 | 0 | gated | $0.55 | $0.55 | $3.30 | chat,fast,vision | openai-chat-completions |
| Leftermute 0x388b…5389 | 0 | gated | $0.2727 | $0.2727 | $1.6362 | chat,coding,json,tools | openai-chat-completions |
| Apex TEE Test 0xe672…7955 | 0 | gated | $0.0926 | $0.0926 | $0.5554 | chat,fast,vision,multimodal,reasoning,long-context,agents | openai-chat-completions |
"Best price" and the seller table are live AntSeed catalog data (advertised $/1M tokens — or $ per generated image for unit-billed image models — not settled amounts). Reputation = on-chain trust (0-100). "Routing" = the SDK's default buyer routing (what the VPR desktop app ships with): a trust ≥ 60 gate on the effective reputation, then cheapest-first among routable sellers; live failover state (per-peer cooldowns) is buyer-side runtime and not included. Model knowledge (TLDR, provider, About) via the VeniceStats enrichment layer. Advertised catalog, not the model used in any specific purchase. "Usage on AntSeed" counts only settlements whose buyers share the per-model split on-chain (metadata v2/v3, opt-in), so every usage figure is a lower bound.