Two of the biggest AI labs made moves within 90 minutes of each other this week. OpenAI shipped GPT-6 Sol and GPT-6 Luna at half the price of its previous 5.6-series models, and Anthropic released Claude Opus 5.5, cutting its own Opus pricing by roughly 40%. Neither company scheduled around the other publicly, but the timing made the contrast hard to miss.
What OpenAI Changed
GPT-6 Sol and Luna are priced at $2 and $10 per million input/output tokens respectively — half of what the equivalent GPT-5.6 tier cost. Sol is positioned for complex coding work, while Luna targets high-volume, clerical-style tasks. OpenAI says Sol makes about half as many factual errors as its predecessor on internal evaluations, putting it close to the reliability of OpenAI's flagship Astra model, at a fraction of the cost.
What Anthropic Changed
Per Anthropic's own announcement, Claude Opus 5.5 costs 40% less to run than Opus 5 on typical workloads — $4 per million input tokens and $20 per million output tokens, with cache reads down to $0.20 per million tokens (a 60% cut). Anthropic also says Opus 5.5 generates output more than 30% faster than Opus 5, and reports it leads on several coding and computer-use benchmarks, including a claimed 66.4% on Terminal-Bench 4.0 versus 52.3% for Opus 5. Anthropic is also raising the five-hour usage limits on its Pro, Max, Team, and Enterprise plans as part of the release.
Why the Timing Matters
Neither company is discounting because AI got cheaper to run out of goodwill — the price cuts reflect real gains in serving efficiency, and they're arriving in the same week because the competitive pressure between frontier labs is now measured in days, not quarters. For any business evaluating which model to build on, that's the practical takeaway: pricing on top-tier models is dropping fast enough that a cost comparison done three months ago is probably already stale.
What This Means If You're Building With AI
For teams building AI-powered products — whether that's a coding assistant, a customer-facing chatbot, or a voice agent — this round of cuts makes frontier-level intelligence noticeably cheaper to put into production than it was even a month ago. The practical move isn't picking a permanent favorite; it's re-checking pricing and benchmarks on a real, recurring cadence before locking in a model for a new build, since the field is genuinely moving week to week right now.
Frequently Asked Questions
Is GPT-6 Sol cheaper than Claude Opus 5.5?
On raw published pricing, GPT-6 Sol ($2/$10 per million input/output tokens) is cheaper than Claude Opus 5.5 ($4/$20). Which is more cost-effective for a specific task depends on how many tokens each model needs to complete it, not just the per-token price.
What is GPT-6 Luna for?
Per OpenAI, Luna is aimed at high-volume, clerical-style work rather than complex reasoning or coding, which is what Sol is positioned for.
Did Anthropic release Opus 5.5 in response to OpenAI?
There's no official confirmation of coordinated timing from either company; the releases landing roughly 90 minutes apart reflects how tightly the two labs are now competing, not a stated response to one another.
