Google: Gemini 3.8 Flash

Gemini 3.8 Flash is a Google model that accepts text, image, video, audio, and file inputs, which makes it suitable for mixed-media workloads without a separate preprocessing step. It offers a context window of about 1.05 million tokens and can return up to 65,536 tokens in a single completion. It supports tool calling and reasoning. The available facts do not confirm structured output support or open weights, so readers who need either should check before committing. Pricing is $0.825 per million input tokens and $4.125 per million output tokens. That makes it a reasonable shortlist candidate for teams processing long documents or multimodal inputs where input volume dominates cost. Its blended benchmark score is 67.8, but that figure comes from only one independent benchmark. The model's standing is therefore thinly evidenced, and buyers should run their own task-specific evaluations before treating it as a proven choice.