OpenAI Now Has Two "Pro" Plans. We Untangled Which One You Should Pay For.

On April 9, 2026, OpenAI quietly launched a second “Pro” tier at $100/month with identical model access to the $200 plan but one-fourth the usage limits. Codex just hit 3 million weekly users and OpenAI saw the gap. The question is whether the new tier closes it or just creates new confusion.

Two Plans Called “Pro”: How We Got Here

For three years, ChatGPT subscribers faced a binary choice: $20 Plus, or $200 Pro. No middle ground. Developers hitting Plus limits hit a wall—the jump to $200 was a hard ask for people who just needed more Codex or a bigger context window, not unlimited everything.

OpenAI’s own team admitted the naming issue straight up: “Naming has never been a strength for the ChatGPT company.” Two plans, same name, different tiers. Confusing as intended? No. A gap they finally decided to fill? Yes.

What $100 Pro Actually Buys You

The $100 plan is not a Plus upgrade. It’s a legitimately different tier. Here’s the breakdown:

The Codex bump is the headline. Through May 31, the $100 plan gets 10x Codex usage (not the permanent 5x). After that promo expires, you drop to 5x. Still solid, but the temporary boost matters if you’re field-testing the tier.

The context window doubling (128K to 256K) is real but not transformative for most people. Deep Research access jumps meaningfully—from about twice a week on Plus to over ten times a week on the $100 tier, which unlocks genuinely research-heavy workflows.

What $200 Pro Buys That $100 Doesn’t

The $200 tier makes sense if you hit one or more of these walls:

• You need the 1M token context window (load full codebases, sprawling research docs, long conversation threads without hallucination creep)
• You use Deep Research daily or running multiple searches in parallel
• You need o3-pro unlimited, not gated access
• Codex is your actual livelihood and 20x is still tight

The jump from 5x to 20x Codex is significant but won’t matter if you’re coding three hours a day. It matters if you’re context-switching between projects or running automated code-generation workflows.

The Claude Max Problem (Why This Tier Exists)

Anthropic’s Claude Max sits at $100/month and has carved out real mindshare among developers. Claude Code crossed a $2.5 billion annualized run rate by early 2026—it’s the fastest-growing piece of Anthropic’s business.

OpenAI can’t ignore that. The $100 Pro directly mirrors Claude Max’s price point and targets the same audience. Same model access, different usage ceiling. OpenAI framed it as closing the gap for heavy users who weren’t ready for $200.

The question now is whether Plus vs $100 Pro vs Claude Max matters anymore or if we’re entering a “pick your LLM brand, then optimize pricing within it” era. Our deeper Claude comparison walks through that decision. If you’re already in the ChatGPT ecosystem, $100 Pro is the natural upgrade.

Our Decision Framework: Plus / $100 / $200

If you use ChatGPT 2–4 hours/week and don’t hit limits: Stay on Plus. Three years at $20 and it still includes thinking, Sora, and Deep Research. No upgrade needed.

If you hit Plus limits weekly but don’t need 1M context or unlimited o3: $100 Pro is the move. The 5x (or 10x temporary) Codex bump handles power-user friction without paying for features you don’t touch.

If Codex is production code, not experimentation: You’re either $100 (if you’ve optimized your prompts and don’t parallelize requests) or $200 (if you do). Codex spend is the real variable here. Check your actual API usage alongside chatgpt.com metrics before deciding.

If you’re comparing to Claude Max: Claude Max locks you into Claude’s codebase—great if Claude Code fits your stack better, costly if you’re switching back and forth. ChatGPT $100 Pro gives you more flexibility because you’re not locked into Deep Research or o-series reasoning; you can dial those down if you don’t use them.

If you need the 1M context window: $200 Pro is the only option. Period. 256K doesn’t cover serious codebase work or truly long conversation threading.

The Promo Trap to Watch

OpenAI’s offering 10x Codex on the $100 plan through May 31. After that, it’s 5x. Some people are upgrading for the temporary boost expecting to downgrade June 1. Don’t do that.

The promo isn’t a bait—it’s a limited-time test period. If you upgrade now and find that 10x is what you actually need for your workflow, downgrading to 5x will sting. Upgrade only if 5x is enough and you’re treating the 10x as bonus breathing room.

Bottom Line

Plus: Still the default. $20 is unchanged for three years; new features keep landing.

Pro $100: The real launch here. It closes a gap that frustrated developers for too long. If Plus feels tight weekly, this is the upgrade.

Pro $200: For people running production code or research at scale, or both. The 1M context window is the differentiator; the Codex ceiling is confirmation you belong here.

Naming is still confusing (thanks for nothing, OpenAI), but the tiers themselves are now rational. Pick the one where your actual usage lives, not the one with the fanciest feature you don’t use.