Perplexity Agent
CoverageRelease Notes
Gemini 3 Flash is available in public preview to Copilot Enterprise, Copilot Business, Copilot Pro, and Copilot Pro+. You'll now be able to access the model...
$10 off the fastest DeepSeek V4.1 Flash, Kimi K3 and GLM 5.3 from Synthetic
Model details
Gemini 3 Flash Preview
Gemini the listed price Flash Preview sits in the Gemini Flash family as a speed- and value-oriented thinking variant intended for agentic workflows, multi-turn chat, and coding assistance. It is positioned as delivering reasoning and tool-use quality close to the larger Gemini Pro models while keeping latency substantially lower, making it well suited to interactive development, long-running agent loops, and collaborative coding sessions where responsiveness matters as much as raw capability. Compared with its predecessor Gemini 2.5 Flash, it is described as offering broad quality improvements across reasoning, multimodal understanding, and reliability, reflecting a generational step rather than a narrow tuning change.
In practical terms, Gemini the listed price Flash Preview is built for developers who want strong reasoning and agentic behavior without paying for frontier-scale latency. It supports a one-million-token context window and accepts text, image, audio, video, and PDF inputs while producing text output, which lets a single model handle mixed-media grounding, long document reasoning, and tool-augmented chat in one pass. Developers can tune the depth of inference through configurable thinking levels (minimal, low, medium, high), and the model exposes structured output, tool use, and automatic context caching so that agent pipelines can keep working state, call external functions, and reuse prior context efficiently across multi-step tasks.
Perplexity Agentgoogle/gemini-3-flash-previewgemini-flash
Quick Info
Powered by- Provider
- Perplexity Agent
- Model key
- google/gemini-3-flash-preview
- Release date
- Dec 17, 2025
- Last updated
- Dec 17, 2025
- Knowledge cutoff
- 2025-01
- Input modalities
- Output modalities
- Capabilities
Cost
- Input token cost
- $0.50
- Output token cost
- $3.00
Limits
- Output tokens
- 65,536 tokens
- Context window
- 1,048,576 tokens
Transparent token rates
Compare Gemini 3 Flash Preview pricing
Rates are shown per one million tokens. Combined means one million input plus one million output tokens.
Latest news about Gemini 3 Flash Preview
Abacus
Coverage
TechCrunch's December 17, 2025 coverage confirmed Gemini 3 Flash launched as the new default in the Gemini app, replacing Gemini 2.5 Flash globally, with users retaining the option to select Pro for math and coding. The model improved multimodal understanding, query intent handling, and visual answer generation across images, audio, and text inputs. Benchmark figures from the report show Gemini 3 Flash scoring 33.7% without tools on Humanity's Last Exam, trailing Gemini 3 Pro's 37.5% and GPT-5.2's 34.5% but well ahead of Gemini 2.5 Flash's 11%. On MMMU-Pro it led all tested models with 81.2%, positioning it as a competitive fast-tier model against frontier alternatives.
Abacus
Official source
A companion Google blog by Josh Woodward, VP of Google Labs and Gemini, framed Gemini 3 Flash as a major everyday-AI upgrade delivering "PhD-level reasoning" at high speed and a leap in multimodal understanding across images, audio, video, and text. It rolled out in the Gemini app on December 17, 2025, with "Fast" for quick answers and "Thinking" for complex problems. The post clarified that Gemini 3 Pro remains available via the model picker and is still recommended for advanced math and code tasks. The positioning underscored Google's intent to close the capability gap between its fast and large model tiers with the Gemini 3 Flash release.
Abacus
Official sourceBenchmark
Google announced Gemini 3 Flash on December 17, 2025, describing it as a fast, cost-effective model offering "Pro-grade reasoning at Flash-level speed" for coding, complex analysis, and interactive apps. The model became the default in the Gemini app and AI Mode in Search, while developers gained access via the Gemini API, Google AI Studio, Vertex AI, Gemini CLI, Android Studio, and Gemini Enterprise. The launch expanded the Gemini 3 family, following Gemini 3 Pro and Gemini 3 Deep Think, with Google reporting over one trillion tokens per day on the API since the prior release. Gemini 3 Flash is positioned as a low-cost entry point to Gemini 3's next-generation reasoning, multimodal, and agentic capabilities across Google's consumer and developer surfaces.
Perplexity Agent
CoveragePreview
Gemini 3 Flash Preview by Vertex AI: pricing, cached input cost, output cost, context window, and capability support. Explore the full generated model page on Sim.