GPT-5.6 Is Here, But You Can't Use It Yet: Limited Preview, Pricing, Full Rollout Coming
OpenAI launches GPT-5.6 limited preview with Sol, Terra, Luna models. Pricing: Sol $5/$30, Terra $2.5/$15, Luna $1/$6 per 1M tokens. Enterprise-only for now, coming to all users in weeks.
Published 84 days ago. Content may be outdated.
Don’t Get Fooled by the Headlines
OpenAI has released the GPT-5.6 series—but this isn’t a full launch. It’s a limited preview.
Here’s what OpenAI officially says:
“During the preview, Sol, Terra, and Luna are available through the OpenAI API and Codex to a limited group of trusted partners and organizations. GPT-5.6 is not available in ChatGPT during the preview.”
Translation: Only enterprises can use it right now. Individual developers? Not yet.
Here’s the Pricing
According to OpenAI’s official documentation:
| Model | Model ID | Input Price / 1M tokens | Output Price / 1M tokens |
|---|---|---|---|
| GPT-5.6 Sol | gpt-5-6-sol | $5.00 | $30.00 |
| GPT-5.6 Terra | gpt-5-6-terra | $2.50 | $15.00 |
| GPT-5.6 Luna | gpt-5-6-luna | $1.00 | $6.00 |
What’s the Difference Between Them?
Sol: The most capable model, designed for complex coding and cybersecurity research
Terra: The balanced option, suitable for most business applications
Luna: The cheapest and fastest, good for everyday text processing tasks
Luna’s $1/1M tokens is pretty competitive—it’s among the lowest in the industry, which puts pressure on open-source inference solutions.
Prompt Caching Got Better
GPT-5.6’s prompt caching improvements include:
- Explicit cache breakpoints — Developers get more control over what gets cached
- 30-minute minimum cache lifespan — Cached content won’t expire quickly
- Updated cache billing:
- Cache writes billed at 1.25x uncached input rate
- Cache reads get 90% discount
For applications that repeatedly call the same System Prompt (like Agents or RAG systems), this could cut costs by 70-80%.
How Big Is the Context Window?
OpenAI’s documentation doesn’t specify it. You might see “1.5M tokens” online, but I couldn’t find that confirmed on their official site. We’ll probably get those details in the technical documentation.
When Can You Use It?
Right Now (Preview Phase)
- ✅ Available through OpenAI API and Codex
- ❌ Not available in ChatGPT (not even for Pro users)
- Only invited enterprises and organizations can participate
Why the Restriction?
OpenAI says it’s a result of coordination with the U.S. government:
“As part of our ongoing engagement with the U.S. government, we previewed our plans and the models’ capabilities ahead of launch. Working in coordination with the government, we are starting with a limited group of trusted partners…”
When Will It Be Available to Everyone?
OpenAI says “in the coming weeks”—no specific date yet.
Any Performance Benchmarks?
OpenAI’s official documentation doesn’t include detailed performance comparisons with other models. We’ll probably see more data once the models reach general availability.
What Developers Need to Know
- Cost — Luna is already at competitive pricing levels
- Availability — Individual developers can’t access it yet; need to wait for general release
- Caching — Using prompt caching can significantly reduce costs
- Timeline — Full rollout expected in weeks, so no need to rush architectural changes now
What Does This Mean for Open-Source Models?
- Price Pressure — Luna’s pricing creates a floor that’s hard for open-source to compete with
- Market Strategy — OpenAI’s phased rollout (enterprises first, then everyone) reduces market friction
- Future Direction — Open-source model providers will likely focus on specialized domains or local deployment
Sources:
- OpenAI Official Help Documentation — Pricing, availability, caching details
- OpenAI Launch Announcement — Official release notes
More Articles
