Contains affiliate links. We may earn a commission — it never changes the verdict.Details
The Stack / Chat Assistants
ChatGPT
02 DecisionsUseful, not essential — if Claude holds your thinking seat, ChatGPT earns its place through breadth: voice, apps, and GPT-5.6 Sol's strong general performance. The $8 Go tier makes it cheap to keep on the bench. Don't pay for both Pro tiers and Claude Max; pick one deep seat.
The job in the stack
The generalist seat: quick answers, brainstorming, voice mode, and the widest plugin/app ecosystem when you need breadth over depth.
Pricing
| Plan | Price | Notes |
|---|---|---|
| Free | $0 | GPT-5.6 Luna, ad-supported, unlimited text chats (uploads, images, and tools still metered) |
| Go | $8/mo | Ad-supported in the US; same GPT-5.6 Luna default and unlimited text chats as Free, with higher limits on everything else |
| Plus | $20/mo | GPT-5.6 Sol access (rolled out from GPT-5.5 the week of July 9), standard limits |
| Pro | $100–200/mo | $100 middle tier added April 2026; $200 remains the ceiling |
| Business | $25/user/mo | $30 monthly; SSO, admin console, SOC 2 |
Pricing verified 2026-08-23. Plans change — the source of truth is ChatGPT's own pricing page.
Models & versions
- GPT-5.6 SolTop model in regular chat on Plus and up, since the week of July 9, 2026 — succeeded GPT-5.5 as the flagship. Retuned August 6 and now powers both Instant and deep reasoning for Plus and Pro
- GPT-5.6 Luna (Free / Go default)Became the Free and Go default on August 6, 2026, replacing GPT-5.5 Instant, with unlimited text chats and a per-message Think button rolling out the week of August 10
- GPT-5.6 (Sol / Terra / Luna)Full family opened via API July 9, 2026. Sol flagship API pricing cut again Aug 21, 2026 — $4/M input, $20/M output (from $5/$30), a promotional rate through Nov 21 aimed at Anthropic and Chinese-model competition. July 30 cuts still stand: Terra $2/$12, Luna $0.20/$1.20. Powers ChatGPT Work and Codex; Free/Go get Terra in those agent surfaces even without a paid plan
The honest read
What works
- +Largest ecosystem: apps, custom GPTs, integrations, voice mode
- +Go tier at $8 is the cheapest paid frontier-model access anywhere
- +GPT-5.6 Sol is a genuinely strong generalist across text, vision, and code
- +ChatGPT Work (new, GPT-5.6 + Codex) runs multi-hour autonomous projects across Drive, Slack, Salesforce, and a growing plugins directory
What breaks
- −Ads on Free and Go tiers — your context is the product
- −Pricing ladder keeps shifting; OpenAI has floated ending unlimited plans
- −Long-document writing quality still trails Claude for operator work
Use it for
- →Operators who want one cheap generalist on the bench
- →Voice-first workflows and quick mobile capture
- →Teams already inside the OpenAI ecosystem
Skip it if
- ×You already pay for Claude Pro and your work is writing/reasoning-heavy
- ×Ad-supported AI in your workflow crosses your trust boundary
Questions operators ask
?Is the $8 ChatGPT Go plan worth it?
As a bench seat, yes — it's the cheapest way to keep a frontier model available. Know that it's ad-supported: if client or business context flows through it, that trade-off is yours to make deliberately.
?ChatGPT or Claude for business writing?
Claude, in our testing and daily use. ChatGPT wins on breadth — voice, apps, images in one place. See our full comparison for the workflow-by-workflow call.
What changed
- News scan: OpenAI cut GPT-5.6 Sol API pricing again on Aug 21 — input $5→$4, output $30→$20 per M tokens, a promotional rate through Nov 21, explicitly framed as competing with Anthropic and Chinese models. Consumer ChatGPT plan prices unchanged.
- News scan: OpenAI moved the Free and Go tiers onto GPT-5.6 Luna on August 6, 2026 and removed the text-message cap — the first time free ChatGPT has been uncapped. Uploads, images, and tools stay metered. A Think button lands for free users the week of August 10. Plus and Pro got a retuned GPT-5.6 Sol that now drives both Instant and deep reasoning. No price moved.
- Re-verified. No change: plan ladder holds ($0 Free / $8 Go / $20 Plus / $100–200 Pro / $25 per seat Business), GPT-5.6 Sol remains the top model on paid plans with GPT-5.5 Instant the everyday default, and the July 30 Terra/Luna API cuts already recorded here are still the current rates.
- NEWS SCAN: OpenAI cut GPT-5.6 API pricing July 30 — Luna dropped 80% to $0.20/$1.20 per M tokens, Terra 20% to $2/$12; Sol unchanged, consumer ChatGPT plans unchanged. Fast mode (2x standard price, ~2.5x speed) replaced Priority Processing. If an automation routes OpenAI API calls on cost, Luna's new floor re-opens that math.
- NEWS SCAN: confirmed the GPT-5.6 chat rollout landed — Plus and up now get GPT-5.6 Sol by default in regular chat (not just Work/Codex via API), and Free/Go moved up to GPT-5.5 Instant. Updated plan notes and models list accordingly.
- NEWS SCAN: OpenAI launched ChatGPT Work (July 9) — an agentic product built on GPT-5.6 and Codex that handles multi-hour autonomous projects across third-party apps via a new Unified Plugins Directory (Google Drive, Slack, Salesforce, GitHub, and more). Rolling out gradually on web/mobile by plan tier; available now via the desktop app, which also absorbed the standalone Codex app.
- NEWS SCAN: GPT-5.6 (Sol/Terra/Luna) opened to the public via API today, ending the ~20-org government-vetted preview that started June 26. ChatGPT app integration timing not yet confirmed — GPT-5.5 remains the top model on paid ChatGPT plans as of this verification.
- Verified plan ladder: Free, Go $8, Plus $20, Pro $100/$200, Business $25–30/user. GPT-5.5 is the top model; GPT-5.6 previewed.
- GPT-5.5 launched across paid plans.
- New $100 Pro middle tier introduced.
Give ChatGPT a job
The generalist seat: quick answers, brainstorming, voice mode, and the widest plugin/app ecosystem when you need breadth over depth.
Try ChatGPT →Alternatives with different jobs
Claude
02 DecisionsThe thinking seat: long documents, hard reasoning, code, and anything where the words have to be right the first time.
Google Gemini
02 DecisionsThe Google-side seat: massive context windows, Deep Research reports, and native reach into Gmail, Docs, and Drive.
Perplexity
01 IntelligenceThe research seat: current-events questions, competitive research, and any claim you need sourced before it ships.