AI Price War Goes Real: GPT-6 and Opus 5.5 Land Same Day

OpenAI released GPT-6 Sol and Luna with a 50% price cut versus GPT-5.6 pricing, with GPT-6 Luna landing at $0.10 per million input tokens and $0.50 per million output. Anthropic countered with Claude Opus 5.5 at roughly half the cost-per-task of Opus 5. The timing was not lost on anyone in the threads, with commenters noting OpenAI almost certainly coordinated the release to collide with Anthropic's launch day.

This is the key bit: token pricing is approaching a level where the cost of inference is becoming irrelevant for many workloads. The thread 'Tokens Too Cheap to Meter' captures the mood. Builders who have been rationing AI calls for cost reasons are watching that constraint evaporate in real time.

The counterpoint in the threads is quality. Several commenters noted that Opus 5.5 produced only marginally better code than its predecessor, and the writing style complaints from Opus 5 apparently persist. Price matters less if the model still makes the same class of errors.