ChatGPT Just Became Free & Unlimited — You Don’t Need a Subscription

Main claim / update

• OpenAI made ChatGPT free and unlimited for regular text chats, removing prior trial/subscription-style constraints for that usage mode.
• The video tests the update live on a free account (no bank card, no subscription).
• However, the “unlimited” upgrade is not fully universal: file uploads, image generation, and other non-text features still have separate limits.

Key technical/product changes demonstrated

“Think” feature moved into the free plan

• Previously paid (shown as locked with a lock icon).
• In free, it appears near the Send button.
• The “Think” mode:
  • Takes longer (example: ~30 seconds vs ~2 seconds for normal replies),
  • Produces more careful, multi-step reasoning,
  • Can reveal assumptions or hidden errors in complex questions.

No model “downgrade” during long chats (text)

• The host compares behavior to the old free tier, which would switch to a weaker model after several messages.
• In the new free tier:
  • No limit/downgrade prompt appeared during a long (~40-message) conversation,
  • Context is maintained across many turns.

Default model behavior

• The host states the free account starts with “GPT 5.6 Luna” immediately after registration (no need to select).
• A higher-end model (“GPT 5.6 Saul”) is described as not available on free—the “most powerful” model remains paid.

What “unlimited free text” covers (examples given)

The host argues the free tier is sufficient for most people’s main ChatGPT use cases, including:

• Writing/editing text
• Studying until concepts are understood
• Debugging code
• Planning trips (even with changing plans)
• Long drafting/restructuring and multi-constraint tasks (when using Think appropriately)

Boundaries and caveats (important)

• Unlimited applies only to text chats

  • File uploads still have limits.
  • Image generation has its own limits (waiting timers appear after a couple attempts).
  • Other modalities and features have separate rules.

• Other features unaffected

  • The host claims items like:
    • Voice conversations
    • Memory between chats
    • Custom GPTs
    • Deeper exploration modes are not changed by the “unlimited text” announcement (still subject to their own limits).

“Safety nets” / abuse handling

• OpenAI still enforces safeguards:
  • “Attempts to bypass them at machine speed” or entering suspicious data can trigger a block.
• The host claims normal users with ordinary questions won’t hit these limits.

Review/guide-like decision framework (how to decide if Plus is worth it)

The video proposes a practical way to evaluate whether paying $20/month is still useful:

1. Check your ChatGPT history from the last ~2 weeks.
2. Count how many chats were pure text (typed + replied; no generation/loading of extra modalities).
3. If most of your usage is that kind of text work (host reports ~4 out of 5 chats), free now covers it completely.
4. Use Plus only if your remaining usage includes:
   • Heavy file work
   • Large-scale image generation
   • Other higher-limit tasks

It also advises:

• If you subscribe and want to test, unsubscribe decisions are reversible; features remain available until the billing period ends.

Ads as the hidden tradeoff (major analysis in the video)

• The host says OpenAI is monetizing the “free unlimited” tier through advertising.
• On Aug 18, OpenAI expanded ads in free ChatGPT to 31 European countries (starting Aug 24).

• Key points:

  • Plus and Pro remain ad-free.
  • Free users in those markets get a variant with ads (and sometimes fewer limits).
  • The “free” menu is described as effectively 3 options:
    1. Unlimited chats with ads
    2. Unlimited/limited text without ads but with restrictions
    3. Paid plan that includes both (unlimited-like access + no ads)

• The host frames this as a product incentive shift:

  • Once ads are tied to the chat interface, incentives may influence the interface/experience over time.
  • Practical takeaway: users in ad-supported regions should expect ads to appear and potentially become more noticeable.

Mentioned alternatives / comparison

• The host briefly contrasts other labs’ free tiers:
  • Gemini: more generous daily limits in the app.
  • Claude: stricter, message-window limits (then you wait).
  • No chat-window ads are claimed for those services.
• The host argues OpenAI’s ad-based free tier differs significantly from others’ approaches.

Sponsor / workshop referenced

• The video sponsor (Out Skuul) is a 3-hour live AI tools workshop:
  • Scheduled this Saturday at 10:00 AM Eastern Time
  • Claimed to have 12M+ attendees completed
  • Teaches choosing the right tool per task and building “agents” that connect systems (e.g., spreadsheets/Notion).
  • Includes bonus sessions, a prompt library, and an AI tool builder (claimed value: $5,000).
• This section functions as a product/tutorial promotion alongside the ChatGPT evaluation.

External product mention: “AI Master” (alternative paid service)

• The host mentions AI Master:
  • Models chosen per task, not per subscription
  • Token-based payments rather than separate monthly subscriptions
  • Mentions image/audio/video generation and “stable characters” (create once, consistent across generations)
  • Mentions a platform that can share work in a feed and support monetization

Main speakers / sources

• Primary speaker: the video host (speaks as “I,” conducting live tests on a free ChatGPT account).
• Official source referenced: OpenAI (announcements about “free/unlimited” and ad expansion).
• Sponsor/source referenced: Out Skuul (AI tools workshop).
• Other referenced companies: Google (Gemini), Anthropic (Claude), xAI (Grok).