Google DeepMind
Build with Gemini
Try Gemini

Our most intelligent workhorse model yet for coding and agents.

Intelligence in a Flash

Advanced reasoning at Flash-level latency and scale.

A minimalist 3D cube icon with a central circle inside a white rounded square on a soft blue glowing background, representing token efficiency.A minimalist white outline of a 3D cube with a central circle inside a dark rounded square, centered on a soft blue glowing background, representing token efficiency.

Versatility across agentic tasks

Optimized for high performance across software engineering, long horizon and knowledge work tasks.

A minimalist lightning bolt icon inside a white rounded square on a soft blue glowing background, representing Gemini Flash.A minimalist white outline of a lightning bolt inside a dark rounded square, centered on a soft blue glowing background, representing Gemini Flash.

Smart and rigorous

Rigorous reasoning efforts for better quality output.

A minimalist stacked layers icon inside a white rounded square on a soft blue glowing background, representing multimodal understanding.A minimalist white outline of a stacked layers icon inside a dark rounded square on a soft blue glowing background, representing multimodal understanding.

Reliable in agentic execution

Navigates roadblocks and resolves coding and real world issues with accuracy.

A minimalist outline of a hand pointing and an eye inside a white rounded square on a soft blue glowing background, representing UI control.A minimalist white outline of a hand pointing and an eye inside a dark rounded square on a soft blue glowing background, representing UI control.

Truly multimodal

Multimodal understanding across text, audio, images, code, and video.

Hands-on

Here are a few ways you can use 3.8 Flash’s improved agentic coding and knowledge work capabilities.

Game built using a looping instruction in Google Antigravity

Gemini 3.8 Flash built this game with a simple prompt using a looping instruction in Google Antigravity. The game uses puzzles, environmental storytelling, and textures generated with Nano Banana to create an immersive 3D level in which you play a wizard navigating a castle.

Fully functional DOS version of Google Maps

Gemini 3.8 Flash builds a fully functional DOS version of Google Maps in a single prompt in Google Antigravity that is fully playable with locations, directions, and Street View.

Topographic map of famous geographical sites

Explore realtime cross-sections, 2D projections, scientific explanations in a topographic map of famous geographical sites built with Gemini 3.8 Flash in Google Antigravity using real datasets from the U.S. Geological Survey and Openweather.

Interactive 3D visualizer

Hardware Anatomy is an interactive 3D visualizer built with Gemini 3.8 Flash in Google AI Studio that generates realistic Three.js renderings of physically-proportioned teardowns for hardware devices. It automatically decomposes devices into layers you can explode and inspect with a deconstruction slider.

Showcase

“Gemini 3.8 Flash excels at long-running, document-heavy workflows, completing more than three times as many tasks as Gemini 3.7 Flash in our evaluations. We're excited to bring its sustained reasoning capabilities to Glean customers who need to turn complex requests into finished artifacts.”

Thai Tran, AI Product Lead, Glean

“At Loopit, we’re integrating Gemini 3.8 Flash into our AI game creation agent for coding, asset reasoning, and rapid visual validation. What stands out is how well-rounded it is. We see strong performance across these tasks, impressive speed, and consistently polished game visuals, helping creators to iterate in real time and turn ideas into playable games faster.”

Haoyuan Guo, CTO, Loopit

Performance

3.8 Flash delivers better intelligence for complex workflows

A line graph comparing various AI models on the DeepSWE V1.1 benchmark, plotting task performance percentage (y-axis) against average cost per task (x-axis, decreasing left-to-right). Gemini 3.8 Flash is highlighted at the top right, indicating high performance at a very low cost.

DeepSWE

On DeepSWE v1.1 (Long-Horizon Software Engineering) 3.8 Flash outperforms most larger frontier models in autonomously solving complex engineering problems end to end, only at a fraction of the cost.

A bar chart comparing performance on "Vals Finance Agent v2" tasks, showing Gemini 3.8 Flash leading at 61.4%, followed by Gemini 3.7 Flash (59.0%), Claude Opus 5 (58.6%), GPT-5.6 Terra (54.4%), Claude Sonnet 5 (53.9%), and GPT-5.6 Sol (53.8%).A bar chart comparing performance on "Vals Finance Agent v2" tasks, showing Gemini 3.8 Flash leading at 61.4%, followed by Gemini 3.7 Flash (59.0%), Claude Opus 5 (58.6%), GPT-5.6 Terra (54.4%), Claude Sonnet 5 (53.9%), and GPT-5.6 Sol (53.8%).

Vals Finance Agent v2

Gemini 3.8 Flash outperforms 3.7 Flash and other frontier models in benchmarks like Vals Finance Agent V2 and Harvey's Legal Agent Benchmark. 3.8 Flash also achieves a 54.9% on HLE-Verified, demonstrating its ability to handle multi-step reasoning across STEM, humanities, and professional fields.

Bar chart comparing AI performance on Harvey's Legal Agent Benchmark. Gemini 3.8 Flash leads with 10.0%, followed by Gemini 3.7 Flash (8.8%), Claude Opus 5 (6.7%), Claude Sonnet 5 (5.0%), GPT-5.6 Sol (2.5%), and GPT-5.6 Terra (0.8%).Bar chart comparing AI performance on Harvey's Legal Agent Benchmark. Gemini 3.8 Flash leads with 10.0%, followed by Gemini 3.7 Flash (8.8%), Claude Opus 5 (6.7%), Claude Sonnet 5 (5.0%), GPT-5.6 Sol (2.5%), and GPT-5.6 Terra (0.8%).

Harvey’s Legal Agent Benchmark

3.8 Flash outperforms 3.7 Flash and other frontier models in benchmarks like Vals Finance Agent V2 and Harvey's Legal Agent Benchmark. 3.8 Flash also achieves a 54.9% on HLE-Verified, demonstrating its ability to handle multi-step reasoning across STEM, humanities, and professional fields.

A bar chart comparing AI models on HLE-Verified multidisciplinary expert reasoning. Gemini 3.8 Flash leads with 54.9%, followed by GPT-5.6 Sol (54.5%), Claude Opus 5 (54.4%), Gemini 3.7 Flash (53.6%), GPT-5.6 Terra (51.1%), and Claude Sonnet 5 (31.0%).A bar chart comparing AI models on HLE-Verified multidisciplinary expert reasoning. Gemini 3.8 Flash leads with 54.9%, followed by GPT-5.6 Sol (54.5%), Claude Opus 5 (54.4%), Gemini 3.7 Flash (53.6%), GPT-5.6 Terra (51.1%), and Claude Sonnet 5 (31.0%).

HLE-Verified

3.8 Flash outperforms 3.7 Flash and other frontier models in benchmarks like Vals Finance Agent V2 and Harvey's Legal Agent Benchmark. 3.8 Flash also achieves a 54.9% on HLE-Verified, demonstrating its ability to handle multi-step reasoning across STEM, humanities, and professional fields.

Model information

Name
3.8 Flash
Status
General availability
Input
  • Text
  • Image
  • Video
  • Audio
  • PDF
Output
  • Text
Input tokens
1M
Output tokens
64k
Tool use
  • Function calling
  • Search as a tool
  • Computer use
Best for
  • Everyday tasks
  • Agentic coding
  • Advanced reasoning
  • Multimodal understanding
  • Knowledge work
Availability
  • Gemini App
  • Gemini Enterprise Agent Platform
  • Google AI Studio
  • Gemini API
  • Gemini AI Mode
  • Google Antigravity
Documentation
View developer docs
Model card
View model card