Gemini 3.5 Flash AI Evaluation | LayerLens - Benchmark Results & Performance

Model

Gemini 3.5 Flash by Google

text+image+file+audio+video->text

Free credits included. No credit card required.

Gemini 3.5 Flash is Google's high-efficiency multimodal model, bringing near-Pro level coding and reasoning at Flash-tier cost and speed. It is highly optimized for coding proficiency and parallel agentic execution. The model uses an ultra-sparse mixture-of-experts architecture that delivers frontier intelligence with exceptional speed and cost efficiency.

Evaluations

across 12 benchmarks

Median latency

6.8s

Context length

tokens

Cost

$1.50 · $9

input · output per 1M tokens

Key takeaways

Gemini 3.5 Flash is Google's high-efficiency multimodal model, optimized for coding proficiency and parallel agentic execution.
It uses an ultra-sparse mixture-of-experts (MoE) architecture to balance frontier intelligence with speed and cost efficiency.
The model scored 96.67% accuracy on the AIME 2024 dataset, demonstrating strong performance in specific problem-solving scenarios.
It achieved 92.93% accuracy on a General Purpose Question Answering dataset, showing solid performance across broader query types.
The model is multimodal, supporting text, image, file, audio, and video inputs, and outputs text.

Technical specifications

Context length: 1M tokens
Max output: —
Architecture: MoE
Modalities: text+image+file+audio+video->text
Parameters: ~1200B
Open weights: No
Licence: Proprietary
Regions: USA
Provider: Google
Provider model key: google/gemini-3.5-flash