AI
GoogleのGemini 3.6 Flash、長期間のエンジニアリングタスクでAIエージェントのトークンコストを最大65%削減、3.5 Proも登場予定
Google's Gemini 3.6 Flash model cuts AI agent token costs by up to 65% on long horizon engineering tasks —and 3.5 Pro is on the way
venturebeat2026年7月21日
日本語要約
GoogleのGemini 3.6 Flashモデルは、長期間のエンジニアリングタスクを実行するAIエージェントのトークンコストを最大65%削減しました。Gemini 3.5 Proも登場予定であり、この進歩は、より費用対効果が高く有能なAIエージェントの未来を示唆しています。これは複雑なAIアプリケーションにおける主要なボトルネックに対処し、その採用を加速させる可能性があります。
English Summary
Google's Gemini 3.6 Flash model has demonstrated a reduction of up to 65% in token costs for AI agents performing long-horizon engineering tasks. This advancement, with Gemini 3.5 Pro also on the horizon, suggests a future of more cost-effective and capable AI agents. It addresses a key bottleneck in complex AI applications, potentially accelerating their adoption.