Parameters
Undisclosed
Context Length
1.05M
Modality
Multimodal
Architecture
Undisclosed
License
Proprietary
Release Date
2 Sept 2026
Knowledge Cutoff
Mar 2026
API Pricing (per 1M)
Input: $0.75 · Output: $3.75
Rank
#34
| Benchmark | Score | Rank |
|---|---|---|
General Text | high 1496 | 9 |
Coding Index | high 0.76 medium 0.74 low 0.73 | 14 27 29 |
Reasoning | high 0.89 | 18 |
Agent Arena | high 0.03 | 19 |
Mathematics | high 0.92 | 27 |
Agentic Coding | high 0.54 | 29 |
General | high 0.76 | 30 |
Web Development | high 1584 | 30 |
LiveBench Average | high 0.76 | 31 |
Agentic Index | high 0.40 medium 0.40 low 0.33 | 40 42 55 |
Coding | high 0.72 | 53 |
Data Analysis | high 0.54 | 61 |
Intelligence Index | high 0.30 medium 0.30 low 0.22 | 80 87 148 |
Overall Rank
#34
Coding Rank
#25
Gemini 3.8 Flash is Google's highly intelligent and fast multimodal reasoning model designed for software engineering and agentic workflows. It offers significant performance gains over previous flash models while handling complex multi-step reasoning tasks.
Architecture specifications are undisclosed for proprietary models.
The Gemini 3.8 model family developed by Google.
Assistant
Online