OpenAI: GPT 5.6 Sol price reduction (until at least Nov 21)
developers.openai.com
August 24, 2026
6 min read
61/100
Summary
OpenAI lists API pricing for GPT-5.6 Sol, Terra, and Luna across short- and long-context requests, with separate rates for input, cached input, cache writes, and output. GPT-5.6 Sol is listed at $4 input and $20 output for short context, rising to $8 input and $30 output for long context; Luna is listed at $0.20 input and $1.20 output for short context. Sol promotional pricing is available at least through November 21, 2026. OpenAI renamed Priority processing to Fast mode on July 30, 2026, while continuing to accept both the "priority" and "fast" service-tier values. The pricing page also covers realtime, image, video, transcription, search, container, file-search, and specialized coding services. Sora 2 video generation is listed at $0.10 per second for 720p, while Sora 2 Pro ranges from $0.30 per second at 720p to $0.70 at 1080p. GPT-Transcribe has an estimated cost of $0.0045 per minute, and web search costs $10 per 1,000 calls plus search-content tokens at model rates. OpenAI is winding down its fine-tuning platform: new users cannot access it, while existing users can create training jobs for the coming months. Fine-tuned models remain available for inference until their base models are deprecated.
Key Takeaways
- OpenAI renamed Priority processing to Fast mode on July 30, 2026, and API requests can use either service_tier: "priority" or service_tier: "fast".
- GPT-5.6 Sol promotional pricing remains available at least through November 21, 2026.
- Daybreak aliases currently map daybreak-blue-latest to GPT-5.6 Sol and daybreak-red-latest to GPT-5.6 Cyber; OpenAI says the aliases will move to new Daybreak frontier models as they are released.
- Sora 2 is listed at $0.10 per second for 720p video generation, while Sora 2 Pro is listed from $0.30 per second at 720p to $0.70 per second at 1080p.
- OpenAI has closed its fine-tuning platform to new users, although existing users can create training jobs for the coming months.
What the discussion said
Commenters treated the Sol discount less as a simple sale than as evidence that frontier-model economics are getting brutally competitive. The optimistic read is that capable intelligence is proving easier to reproduce and improve than monopoly-minded forecasts assumed: Chinese labs pretrain serious models of their own, distillation can carry knowledge forward, and API price wars make strong systems available to far more developers. Several users also argued that Sol is already close to, and occasionally better than, Anthropic’s premium Fable on practical coding review, while its lower cost and more concise answers make switching plausible. The thread was not convinced that headline token prices settle the question. Some noted that raw per-token comparisons miss differing token consumption and that the best model can justify a meaningful premium when mistakes are expensive. Others said DeepSeek remains much cheaper for little apparent sacrifice, though opponents insisted current benchmarks put it behind newer competitors. A temporary discount drew the sharpest operational skepticism: production teams cannot comfortably evaluate around a price that may vanish, whereas others argued model selection should be revisited every month or two anyway. Subscription users felt left behind by weekly limits and opaque resets, even as others argued flat-rate access is already heavily subsidized. Underneath the excitement sat concern that lower prices may signal demand pressure, alongside warnings that foreign open-weight models pose trust and data-exfiltration risks for sensitive users.
Where opinion split
The central fight was whether the temporary Sol cut is a genuinely durable win for developers or a poor basis for production adoption. Supporters see a fast-moving market where frequent re-evaluation is normal and lower-cost frontier capability forces every provider to compete; skeptics argue a time-limited rate makes cost planning and meaningful evaluations impossible until pricing is stable.
Community Sentiment
Positives
- Frontier-level reasoning is being pushed toward commodity pricing, weakening the prospect that one lab can lock up intelligence behind an unassailable moat.
- Several practical users report Sol catching coding-solution errors that Fable missed, making the cheaper model a credible alternative rather than merely a budget tier.
- The reduced API rate makes Sol materially more attractive against premium rivals, especially where OpenRouter pricing reaches roughly $2 input and $10 output per million tokens.
- Competition among proprietary APIs and open-weight alternatives is giving developers more leverage, with users openly considering abandoning expensive premium subscriptions.
Concerns
- A discount scheduled to expire gives production teams no dependable cost floor, so some refuse even to run evaluations until the economics are durable.
- Headline token prices can mislead because models consume different numbers of tokens; apparent savings may disappear when a cheaper model is more verbose or less efficient.
- Subscribers still face restrictive, poorly explained weekly caps and rolling resets, leaving flat-fee customers outside the benefit of the API price cut.
- Cheap foreign open-weight models raise trust concerns for corporate, government, and military workloads if hidden behavior could expose sensitive information.
