AI 구독 가격

Unlimited text chats (Aug 2026); reasoning, tools and image gen are capped.

Budget tier; higher limits than Free (~160 messages / 3h).

Flagship model, ~160 messages / 3h, Deep Research, Sora, Agent mode, Codex.

한도: ~160 messages / 3h

Pro-level usage per seat, admin controls, data excluded from training.

5× Plus usage plus the Pro reasoning model.

한도: 5× Plus (~800 messages / 3h)

20× Plus usage; effectively unlimited within fair-use guardrails.

한도: 20× Plus (~3,200 messages / 3h)

Basic access; rolling 5-hour session limit plus a weekly cap.

~45 messages / 5h plus a weekly cap; Projects and Claude Code.

한도: ~45 messages / 5h plus a weekly cap

Team workspace and central billing (min. seats apply).

5× Pro usage (5h + weekly caps); priority access to new models.

한도: 5× Pro (~225 messages / 5h) plus a weekly cap

20× Pro usage; heaviest individual tier.

한도: 20× Pro (~900 messages / 5h) plus a weekly cap

Entry Gemini app tier; available in 160+ countries.

~4× higher usage than Free, video generation, 2 TB storage.

Highest limits, top models, Flow/Whisk credits, 30 TB storage.

Entry paid Grok tier with higher limits than free.

Full Grok with higher rate limits and frontier models.

Top tier with the largest limits and earliest access.

Standard search with a few Pro searches per day.

300+ Pro searches/day, model choice, Labs and file uploads.

Highest limits, early features, top research models.

Free Le Chat access with basic limits.

Higher limits, larger context, more capabilities.

Team workspace with shared library and admin.

Limited completions and chat requests per month.

Unlimited completions, agent mode, premium request quota.

Org policy controls and seat management.

Highest premium-request quota and all models.

3,900 AI credits/user, knowledge bases, priority model access.

한도: 3,900 AI credits / user / month

Free tier with limited agent requests.

$20 of included usage, unlimited Auto, MCPs, skills and hooks.

Per-seat team plan with admin and pooled usage.

~3× Pro usage for heavier agent workflows.

Highest individual limits across frontier models.