OpenAI has reduced pricing for its GPT-5.6 Sol model by 50% through the OpenRouter API gateway, according to the platform's pricing page.

The flagship model now costs $4 per million input tokens and $20 per million output tokens. GPT-5.6 Sol launched in July 2026 with a 1.1 million token context window and knowledge cutoff of February 2026.

OpenAI positions the model as optimized for complex reasoning, coding workflows, and agentic tasks. The company specifically highlights its strength in command-line operations and multi-step coding problems.

Benchmark Performance

GPT-5.6 Sol shows variable performance across different reasoning modes on Artificial Analysis benchmarks. The "Max" configuration achieves a 47.0 Intelligence Index and 94.1% on GPQA Diamond, while the "Low" reasoning mode drops to 33.5 and 89.8% respectively.

Coding performance remains more consistent across configurations, with the Coding Index ranging from 65.1 to 78.3. Terminal-Bench scores vary significantly by difficulty level, from 1.0% on version 4.0 to 89.5% on version 2.1.

The model's agentic capabilities peak at 50.2 on the Agentic Index in Max mode but fall to 28.1 in non-reasoning configuration. Hallucination rates hover between 7.2% and 10.6% across different reasoning settings.

OpenRouter routes requests across multiple hosting providers based on user-selected modes prioritizing price, speed, or accuracy. The platform continuously monitors uptime and switches to backup providers when errors occur.

The price cut positions GPT-5.6 Sol more competitively against other frontier models as OpenAI expands enterprise adoption of its latest generation.