OpenAI has reduced the pricing of two models in its GPT-5.6 family as competition intensifies in the AI market, particularly from lower-cost Chinese open-weight models such as Kimi K3.
OpenAI CEO Sam Altman announced the price cuts on X, revealing that GPT-5.6 Luna and GPT-5.6 Terra will now be significantly cheaper to use.
"Major price cuts today: 80 per cent drop for GPT-5.6 Luna, now $0.20 (around Rs 20) per million input tokens and $1.20 (around Rs 115) per million output tokens," Altman wrote.
He also announced a 20 per cent price reduction for GPT-5.6 Terra, bringing its pricing down to $2 (approximately Rs 191) per million input tokens and $12 (around Rs 1,150) per million output tokens.
The revised pricing makes both Luna and Terra more affordable than Kimi K3, one of the largest open-weight AI models currently available. Kimi K3 is priced at $3 (roughly Rs 287) per million input tokens and $15 (around Rs 1,436) per million output tokens.
Since its launch, Kimi K3 has attracted significant attention for delivering performance comparable to GPT-5.6 Sol and Anthropic's Claude Fable 5 on certain benchmarks while operating at a substantially lower cost.
OpenAI introduces Fast mode for GPT-5.6 Sol
Alongside the price reductions, OpenAI has introduced a new Fast mode for GPT-5.6 Sol, its flagship AI model.
According to Altman, GPT-5.6 Sol's Fast mode delivers responses up to 2.5 times faster while costing twice as much as the standard version, with no compromise in intelligence.
He said the company aims to provide the best balance between cost and performance across its range of AI models.
OpenAI also announced improvements in the operational efficiency of GPT-5.6 Sol. The company said it used the model itself to optimise its serving infrastructure, reducing serving costs by 20 per cent and improving token-generation efficiency by more than 15 per cent.
The pricing changes come as businesses increasingly scrutinise AI spending and seek better returns on investment. As enterprise AI costs have grown substantially, companies including Uber and Walmart have reportedly begun tightening budgets for AI deployments.
At the same time, OpenAI and other US AI companies are facing growing competition from Chinese open-weight models, which can be downloaded, customised and deployed on an organisation's own infrastructure. Kimi K3, for example, is available as a free download, allowing developers to run and fine-tune the model locally.
OpenAI's latest pricing strategy is also expected to intensify competition with rivals such as Anthropic and Google, both of which have recently promoted more affordable AI offerings. Meta and xAI have likewise introduced increasingly cost-efficient models, signalling the possibility of a broader price war among major AI companies competing for enterprise customers.
