OpenAI's $100 Codex Gambit Targets Anthropic's Sweet Spot
Last month, I watched our engineering team burn through ChatGPT Plus limits by Tuesday. Every week. Our choices were painful: stick with $20/month frustration or jump to $200/month overkill for a 12-person team.
OpenAI apparently heard similar complaints. Their new $100/month ChatGPT Pro tier launched April 9th, targeting exactly this gap. But this isn't just pricing optimization—it's a direct shot at Anthropic's market position.
The Numbers Game
Here's what changed:
- Plus ($20): Base Codex limits, rebalanced for weekly steady use
- New Pro ($100): 5x more Codex usage (10x through May 31st promo)
- Enterprise Pro ($200): 20x limits, mysteriously missing from public pricing
The structure now mirrors Anthropic's $20 Pro, $100 Max 5x, $200 Max 20x tiers perfectly. Coincidence? Doubtful.
<> "Industry observers view the move as a strategic play to dominate AI coding and counter Anthropic, with the $100 tier providing a 'crucial middle ground' for professionals."/>
What Developers Actually Get
The $100 tier unlocks GPT 5.4 Pro—what developers describe as "slow but thorough" for research and hard bugs. You get unlimited Instant and Thinking models alongside 5x Codex capacity.
But here's the catch: no plan offers unlimited usage. Even at $200/month.
Hacker News users clarified the real multipliers:
- Plus: 0.3x-1x baseline
- $100 Pro: 1.5x-5x
- $200 Pro: 6x-20x
That promotional 10x boost through May? Classic drug dealer tactics. First taste is always the strongest.
The Anthropic Chess Move
This pricing restructure forces a conversation OpenAI wants: "Why pay for Claude when we match their tiers?"
Before this announcement, developers frustrated with ChatGPT limits naturally migrated to Claude. Now they're stuck choosing between nearly identical pricing structures. The differentiation shifts from cost to capability.
Smart developers are already planning hybrid stacks—Claude 5x plus ChatGPT Codex 5x. The tooling landscape just got more fragmented.
The Real Strategy
OpenAI isn't just filling a pricing gap. They're:
1. Capturing revenue from pros avoiding enterprise sales calls
2. Pressuring Anthropic to compete on features, not pricing
3. Testing demand elasticity at the $100 price point
The missing $200 tier from public pricing pages isn't an oversight—it's intentional friction pushing enterprise prospects toward sales conversations.
Red Flags Worth Watching
Promotional limits create churn risk. Developers who build workflows around 10x capacity will hit walls come June 1st. That's either brilliant upgrade pressure or customer satisfaction disaster.
Rate limit transparency remains murky. Users report seeing $120/month pricing in some interfaces. Variable pricing suggests A/B testing or regional adjustments.
The "rebalanced" Plus plan favors weekly distribution over daily bursts. Translation: we're managing server costs by smoothing demand spikes.
My Bet
This pricing move succeeds in the short term but creates long-term fragmentation. Developers will optimize for multiple providers rather than committing to one. OpenAI captures market share from Anthropic but loses the "obvious choice" positioning.
The real winner? Cursor, Replit, and other coding tools that abstract away the underlying model provider. When pricing structures converge, developer experience becomes the only differentiator.
OpenAI just commoditized their own premium positioning. Smart tactically, questionable strategically.
