OpenAI and Anthropic have implemented substantial price reductions for their large language model (LLM) APIs throughout 2026, with some models seeing cuts of up to 80%. This aggressive pricing shift, noted by developer Max Woolf and confirmed by company announcements and pricing trackers, indicates a maturing market and a focus on broader adoption for advanced AI.OpenAI's GPT-5.6 Luna model experienced an 80% price cut on July 30, 2026, lowering its cost to $0.20 per million input tokens and $1.20 per million output tokens. This reduction positioned Luna competitively against models like DeepSeek V4 Pro, according to an OpenAI chart cited by Woolf. Additionally, OpenAI reduced the price for GPT-5.6 Sol by approximately 20% to $4 per million input tokens and $20 per million output tokens on August 21, a promotional rate set to last for three months. Further reductions arrived on September 22, when OpenAI released GPT-6 Sol and GPT-6 Luna at half the input and output prices of their predecessors, with GPT-6 Sol at $2 per million input tokens and $10 per million output tokens, and GPT-6 Luna at $0.10 per million input tokens and $0.50 per million output tokens.Anthropic also adjusted its pricing structure. Claude Sonnet 5, initially launched with a temporary price of $2 per million input tokens and $10 per million output tokens on June 30, had this introductory rate made permanent on August 10. The company introduced Claude Opus 5 on July 24 at $5 per million input tokens and $25 per million output tokens. On September 1, Anthropic released Fable 5.1, maintaining a price of $10 per million input tokens and $50 per million output tokens but cutting the cache read cost by 75% compared to Fable 5. Cache read costs are a significant component of token expenditure for agentic coding workloads, often accounting for 96% to 99% of the total. On September 22, Anthropic launched Claude Opus 5.5, which the company stated requires less compute to serve than Opus 5, and its pricing reflects this efficiency. Claude Opus 5.5 is priced at $4 per million input tokens and $20 per million output tokens, representing a 20% reduction from Opus 5. On October 7, Anthropic introduced Claude Haiku 5.5, priced at $0.10 per million input tokens and $0.50 per million output tokens for prompts up to 100,000 tokens, a 90% reduction from Haiku 4.5 for such requests. The company also lowered the price of cache reads on Claude Sonnet 5.5 by 50% to $0.10 per million tokens, which is expected to reduce the cost of most agentic tasks on that model by approximately 20%.These price reductions are attributed to several factors, including infrastructure efficiency gains, such as optimized inference infrastructure, custom silicon deployments, and improved model architectures. Economies of scale as usage volumes increased through 2026 have also contributed to the ability of providers to lower costs while maintaining profit margins. The competitive landscape, particularly from open-weight models, is another driving force behind the price adjustments. Analysts suggest that the ability to swap models easily through tools like OpenCode and OpenRouter, combined with the emergence of cost-effective alternatives like DeepSeek V4.1 Flash 1017 and Kimi K3.1 Pro, pressures major LLM providers to lower their rates.The overall trend indicates a shift in the AI market, where advanced language models are becoming more accessible to developers. This commoditization of token pricing means that while individual tokens are cheaper, overall AI spending by companies is surging due to increased usage and experimentation. Companies are advised to implement strategies such as batch processing and prompt caching to manage these rising costs effectively.