OpenAI has slashed the price of its flagship GPT-5.6 Sol model by 20 percent in the API, bringing input costs down from $5 to $4 per million tokens and output costs from $30 to $20 per million tokens. The promotional pricing is available at least through November 21, 2026, according to OpenAI’s updated pricing page.
The move marks the third price adjustment in the GPT-5.6 family since its July 9 launch. OpenAI cut Luna by 80 percent and Terra by 20 percent on July 30, leaving Sol untouched at the time. Now the top-tier model has followed suit, meaning every tier in the GPT-5.6 lineup has seen at least one price reduction within six weeks of going live.
What the new pricing looks like
The revised short-context rates place Sol at $4 input and $20 output per million tokens, Terra at $2 input and $12 output, and Luna at $0.20 input and $1.20 output. Long-context pricing, which applies above 272,000 input tokens, remains at double the input rate and 1.5 times the output rate for Sol.
Cached input tokens for Sol now cost $0.40 per million, down from $0.50, while cache writes are priced at $5 per million. The changes bring Sol closer to the cost territory that Terra occupied at launch, potentially reshaping how enterprises allocate workloads across the three-tier family.
A pricing war across the frontier
The cut arrives amid a broader AI industry race on price. Google released Gemini 3.7 Flash on August 13 at $0.75 input and $3.75 output per million tokens, half of what Gemini 3.6 Flash cost at launch. Meanwhile, DeepSeek moved V4-Pro to general availability on the same day but raised its own prices sharply, introducing peak and off-peak billing with output costs jumping to $3.96 per million during busy hours.
The competing signals suggest the frontier AI market is splitting into two strategies: discount-driven labs chasing developer volume, and proven providers testing how much they can charge once a model demonstrates real-world value. Sol’s new price positions OpenAI firmly in the first camp, at least through November.
For a company running large-scale coding or research workloads through Sol, the reduction translates to meaningful savings. A workflow that processes 100 million input tokens per day now costs $400 instead of $500, a daily saving of $100 that compounds quickly across engineering teams.
discussion