Gemini 3.5 Flash
Google ・ モデルID: gemini-3.5-flash ・ 取得日 2026-07-05
コンテキスト長
1,048.576K tok
最大出力
65.536K tok
入力単価(/100万tok)
$1.5
出力単価(/100万tok)
$9
この仕様でできること(換算目安)
| 項目 | トークン | 日本語文字(目安) | A4ページ(目安) |
|---|---|---|---|
| コンテキスト長(入力に入る最大量) | 1,048,576 | 約70万字 | 約437ページ |
| 最大出力(1回の応答の上限) | 65,536 | 約4.4万字 | 約27ページ |
換算は 1文字 ≈ 1.5トークン・A4 ≈ 1,600字 の目安です。実際のトークン数はトークナイザや書式で変わります。
向いている使い方
100万トークン級の非常に長いコンテキスト。長大な文書群・コードベース全体・長い会話履歴をまとめて渡す用途に向きます。
キャッシュ入力単価
context_cache: $0.15 / storage_per_hour: $1
補足(公式ドキュメントより)
Most intelligent Gemini model for sustained frontier performance on agentic/coding tasks. Context/max output figures based on Gemini 3 family standard limits confirmed for gemini-3.1-pro-preview; not individually re-verified for this exact model on its own spec page.
同じ提供元(Google)の他モデル
| モデル | コンテキスト長 | 最大出力 | 入力 /1M |
|---|---|---|---|
| Gemini 3.1 Pro (Preview) | 1,048.576K | 65.536K | $2 |
| Gemini 3.1 Flash-Lite | — | — | $0.25 |
| Gemini 2.5 Pro | — | — | $1.25 |
| Gemini 2.5 Flash | — | — | $0.3 |
| Gemini 2.5 Flash-Lite | — | — | $0.1 |
出典
https://ai.google.dev/gemini-api/docs/pricing (Google公式 ・ 取得日 2026-07-05)