Gemini 3.8 Flash
MultimodalGemini 3.8 Flash is Google's most capable Flash model, built for long-horizon software engineering, autonomous agents, and complex enterprise workflows. It accepts text, image, video, and audio input, offers a 1M-token context window, and launched on September 2, 2026 at $0.75 per million input tokens.
Key Specifications
Parameters
-
Context
1.0M
Release Date
September 2, 2026
Average Score
59.2%
Timeline
Key dates in the model's history
Announcement
September 2, 2026
Last Update
September 3, 2026
Technical Specifications
Parameters
-
Training Tokens
-
Knowledge Cutoff
March 31, 2026
Family
-
Capabilities
MultimodalZeroEval
Pricing & Availability
Input (per 1M tokens)
$0.75
Output (per 1M tokens)
$3.75
Max Input Tokens
1.0M
Max Output Tokens
65.5K
Supported Features
Function CallingStructured OutputCode ExecutionWeb SearchBatch InferenceFine-tuning
Benchmark Results
Model performance metrics across various tests and benchmarks
Other Tests
Specialized benchmarks
BioMysteryBench
Human difficult; terminal, bioinformatics tools, and restricted internet access. • Self-reported
CharXiv-R
No tools; self-computed. • Self-reported
DeepSWE 1.1
DeepSWE v1.1; mini-swe agent harness; high thinking. • Self-reported
GDP.pdf
All-pass rate; self-computed. • Self-reported
GDPval-AA
GDPval-AA v2 Elo (max_score 3000); sourced from the Artificial Analysis public leaderboard. • Self-reported
Harvey's Legal Agent Benchmark
All-pass rate; sourced from Vals.AI. • Self-reported
HLE-Verified
Full 1,811-item verified set; self-computed. • Self-reported
LABBench2
Biology real-world research tasks with terminal, bioinformatics tools, and internet access. • Self-reported
LVBench
Static; 1024 frames; no tools; self-computed. • Self-reported
OSWorld 2.0
Partial score; batched tool calls enabled; maximum over three runs. • Self-reported
Terminal-Bench 2.1
Terminus 2 default agent harness; self-computed. • Self-reported
Terminal-Bench 4.0
Highest scoring thinking level reported by the Terminal-Bench public leaderboard. • Self-reported
Vals Finance Agent v2
Financial analyst tasks; sourced from Vals.AI. • Self-reported
License & Metadata
License
proprietary
Announcement Date
September 2, 2026
Last Updated
September 3, 2026
Compare Gemini 3.8 Flash
All comparisonsArticles about Gemini 3.8 Flash
Similar Models
All ModelsGemini 2.0 Flash Thinking
MM
Best score:0.7 (GPQA)
Released:Jan 2025
Gemini 2.0 Flash-Lite
MM
Best score:0.5 (GPQA)
Released:Feb 2025
Price:$0.07/1M tokens
Gemini 2.5 Flash-Lite
MM
Best score:0.6 (GPQA)
Released:Jun 2025
Price:$0.10/1M tokens
Gemini 3.5 Flash-Lite
MM
Released:Jul 2026
Price:$0.30/1M tokens
Gemini 3 Pro
MM
Best score:0.9 (GPQA)
Released:Nov 2025
Price:$2.00/1M tokens
Gemini 3.5 Flash
MM
Released:May 2026
Price:$1.50/1M tokens
Gemini 3.1 Pro
MM
Best score:0.9 (GPQA)
Released:Feb 2026
Price:$2.50/1M tokens
Gemini 3.7 Flash
MM
Released:Aug 2026
Price:$0.75/1M tokens
Recommendations are based on similarity of characteristics: developer organization, multimodality, parameter size, and benchmark performance. Choose a model to compare or go to the full catalog to browse all available AI models.