GPT-5.6 Sol, OpenRouter cuts pricing by 50%
Original: GPT-5.6 Sol Pricing Cut by 50%
Why This Matters
A 50% price cut on a flagship reasoning model signals intensifying competition in the enterprise AI API market.
OpenRouter has applied a 50% discount to OpenAI's GPT-5.6 Sol model, reducing input pricing to $2.50/1M tokens and output to $15/1M tokens. The model supports a 1M-token context window and was released July 9, 2026, with a February 2026 knowledge cutoff.
OpenRouter is offering GPT-5.6 Sol — OpenAI's flagship model in the GPT-5.6 series — at a 50% price reduction via its platform. The discounted rate brings input costs to $2.50 per 1M tokens and output to $15 per 1M tokens, down from the standard $5.00 and $30.00 respectively. Cache read pricing is also halved to $0.25/1M tokens. GPT-5.6 Sol is described as optimized for complex reasoning, coding, and agentic workflows, with particular strength in command-line tasks, multi-step coding, and long-horizon problem solving. The model supports a 1M-token context window and was released on July 9, 2026, with a knowledge cutoff of February 2026. Multiple providers serve the model through OpenRouter, including OpenAI directly, Azure, Azure EU, Azure US, and Amazon Bedrock US. Performance benchmarks show GPQA Diamond scores ranging from 90.6% to 92.3% across providers, and TAU-Bench scores between 70.4% and 80.0%. Platform-wide 3-day uptime stands at 99.76% availability.