Breaking

OpenAI and Anthropic Release Lower-Cost Flagship AI Models as Price Competition Intensifies

September 23, 2026 at 12:11 EDT

Foundation Models
AI Agents

OpenAI and Anthropic both introduced cheaper high-end AI models on September 22, sharpening a price-performance contest among frontier model providers while keeping the focus on enterprise and developer workloads. OpenAI rolled out GPT-6 Sol and GPT-6 Luna for ChatGPT Work, Codex and the API, while Anthropic introduced Claude Opus 5.5 with lower token prices, faster generation and higher usage limits for paid plans.

September 22, 2026 · OpenAI and Anthropic

Cheaper frontier models arrive

OpenAI released GPT-6 Sol and GPT-6 Luna, while Anthropic introduced Claude Opus 5.5 with lower token prices, faster output and higher usage limits.

$0.10

GPT-6 Luna input price per 1M tokens

40%

Claude Opus 5.5 typical workload cost reduction

30%+

Claude Opus 5.5 output speed gain over Opus 5

Claude Opus 5.5 cost drop, drawn to scale

100

Opus 5 baseline

Opus 5.5 typical cost

Anthropic says Opus 5.5 costs 40% less to run than Opus 5 on typical workloads.

Why users like it

Lower unit costs make stronger models more practical for coding agents, document workflows and high-volume internal tools.

What still matters

Real cost depends on caching, long-context prompts, plan eligibility, regional processing and retries.

The new model-buying logic

Pick model tier→Reuse cached context→Measure cost per task

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.