Gemini 3.8 Flash
Gemini 3.8 Flash by Google is its most intelligent Flash model, built for long-horizon software engineering, autonomous agents, and enterprise workflows.
Model Overview
Capabilities, design details, and architectural traits
Gemini 3.8 Flash - Google's most intelligent Flash model for long-horizon work
Gemini 3.8 Flash is Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows. It is generally available and ready for production use, and it delivers its gains at the same speed and low cost as the previous 3.7 Flash.
| Trait | Detail |
|---|---|
| Positioning | Most intelligent Flash model in the Gemini lineup, often approaching higher-cost frontier models |
| Built for | Long-horizon software engineering, autonomous agents, and complex enterprise workflows |
| Reasoning control | Tunable thinking levels: low, medium, high, with medium as the default |
| Context | 1M token context window and 64k max output tokens |
| Training edge | Coding and reasoning gains driven in part by rigorous training in the demanding domain of cybersecurity |
Tunable thinking levels
A defining feature of 3.8 Flash is its reasoning control. Developers can set the thinking level to low, medium, or high, trading depth against latency and cost per task. The default is medium.
A cybersecurity-trained sibling
3.8 Flash shares its foundational intelligence with Gemini 3.8 Flash Cyber, a variant focused on vulnerability detection and automated patching. Both variants are accelerated by long-running agentic loops that recursively evaluate and refine the underlying models, and the cybersecurity training behind this shared core is a documented driver of 3.8 Flash's coding and reasoning gains.
Benchmark Performance
Independent evaluations · Artificial Analysis
Accuracy & Capability Details
Compare Models Side-by-Side
Evaluate specifications, pricing, and independent benchmark indices
| Model Details | |||
|---|---|---|---|
| General Info | |||
| Provider | Anthropic | Anthropic | |
| Release Date | September 2, 2026 | September 1, 2026 | July 24, 2026 |
| Knowledge Cutoff | - | - | May 2026 |
| Context & Limits | |||
| Context Window | - | - | 1M |
| Pricing (per 1M tokens) | |||
| Input Pricing | $0.75 Best Input Pricing | $10 | $5 |
| Output Pricing | $3.75 Best Output Pricing | $50 | $25 |
| Modalities | |||
| Inputs | text | textimagefile | textimage |
| Outputs | text | text | text |
| Benchmarks (0-100) | |||
| Intelligence Index | 58.7 | 65.7 Best Intelligence Index | 63.1 |
| Coding Index | 76.3 | 81.6 Best Coding Index | 78.0 |
| Agentic Index | 50.0 | 61.3 Best Agentic Index | 59.2 |
GPQA Benchmark
Graduate-level reasoning and expert Q&A evaluation.
Humanity's Last Exam
Extremely difficult logical reasoning and knowledge.
Long Context Reasoning
Logical reasoning over long context windows.
Independent evaluation data provided by Artificial Analysis. To view the latest benchmarks and full details, visit their official site.