Gemini 3.8 Flash
by Googletext+image+file+audio+video->text6 endpoints
Neura Intelligence Index
78.8/ 100
Rank
#37—
Confidence
lowBenchmarks
1(3% coverage)
Overview
Gemini 3.8 Flash is Google's most intelligent Flash model with significant gains from 3.7 Flash across software engineering, agentic tasks, and multi-step reasoning.
Capabilities
Text generation
Image understanding
Audio processing
Video understanding
Tool use / Function calling
Structured output
Extended reasoning
Prompt caching
Long context
File input
Top-P sampling
Stop sequences
Deterministic seed
Modalities
Input
TextImageVideoFileAudio
Output
Text
Technical Specifications
Context Window
1.0M tokens
Max Output
65.5K tokens
Tokenizer
Gemini
Uptime (24h)
9995.7%
Country
🇺🇸US
Release Date
September 2, 20262w
Supported Parameters
include_reasoningmax_tokensreasoningreasoning_effortresponse_formatseedstopstructured_outputstemperaturetool_choicetoolstop_p
Performance Over Time
Pricing
Per 1M tokens
Input$0.75
Output$3.75
Cache read$0.07
Cache write$0.04
Audio$0.75
Reasoning$3.75
Image$0.75
Added
September 2, 2026
Last synced 9/16/2026