OpenAI and Anthropic slash prices as Chinese rivals gain ground in AI cost war

A price war is heating up between leading US AI labs and their Chinese rivals, as OpenAI and Anthropic slash prices on their mid-tier models to hold onto customers who are increasingly defecting to cheaper alternatives from Chinese developers like Moonshot and DeepSeek.

The moves come as corporate AI users confront rising bills, prompting many to impose usage caps or test lower-cost models. According to Silicon Data's token price index, prices for leading US models have dropped by nearly a quarter since mid-July, reflecting a broader shift as open-source Chinese models narrow the performance gap.

OpenAI cut the price of GPT-5.6 Luna, its fastest and most affordable model, by 80% — from $1 to $0.20 per million input tokens and from $6 to $1.20 per million output tokens. Anthropic launched Claude Opus 5 at half the price of its flagship Fable 5, at $5 per million input tokens and $25 per million output tokens. The company also called off a planned price increase for its Sonnet 5 model.

Defending the top

These cuts apply primarily to mid-tier offerings, making them more competitive with Chinese models. However, headline token prices don't tell the whole story. More capable models can complete tasks using fewer tokens, and "effort" settings allow users to trade computing power for performance, complicating direct comparisons.

Benchmarks from Artificial Analysis show that Anthropic's Opus 5 at medium effort delivers performance and cost per task similar to Moonshot's Kimi K3 at max effort. OpenAI's GPT-5.6 Luna at max effort performs similarly to DeepSeek's V4 Flash, but costs just under twice as much per task.

Despite the cuts, prices for the most advanced models remain flat or rising, according to Mantas Lukauskas, AI tech lead at Hostinger. He called the recent pricing changes the "first real test" of whether US labs can protect their premium offerings: "The US labs have cut the middle and are defending the top."

As OpenAI and Anthropic prepare for IPOs at trillion-dollar valuations, this price war underscores the tension between maintaining massive AI infrastructure spending and proving to investors that the industry can generate returns. For developers and enterprises, the immediate takeaway is clear: the cost of AI is falling, but the trade-offs between price and performance are becoming more nuanced.