Plans & Pricing
When a new model is verified, you get it — automatically. Alongside them runs Opal, our own model in development for silicon and protonic compute.
Everything, every day
A$22.67 / month
A$67.99 billed every 3 months
Leading AI models, plus Opal in preview.
No ceiling
A$176.33 / month
A$528.98 billed every 3 months
Everything in Pro, plus:
Cancel anytime. Fair-usage limits apply. Prices in AUD and exclude tax. Billing opens progressively from the waitlist.
Models marked “Planned” are on the roadmap and switch on only after live verification.
| Feature | Pro | Max |
|---|---|---|
| Essentials | ||
| Daily usage | Standard | 5× Pro |
| Response times | Fast[2] | Fast[2] |
| Chat history | Unlimited[1] | Unlimited[1] |
| Access on web, desktop and iOS | ||
| Memory and context | Standard[2] | Maximum[2] |
| Early access to new features | ||
| Chat | ||
| Opal Instant | ||
| Opal preview (our own model, in development) | Priority | |
| Switch models mid-chat | ||
| GPT-6 Astra | ||
| GPT-5.6 Sol | Planned | Planned |
| GPT-5.6 Luna | ||
| GPT-4o mini | ||
| Claude Opus 5 | Planned | Planned |
| Claude Fable 5 | Planned | Planned |
| Gemini 2.5 Flash Lite | ||
| Gemini 3.1 Pro | Planned | Planned |
| Llama 3.3 70B Instruct | ||
| New models as they come out | ||
| Reasoning effort | High[3] | Maximum[3] |
| Bring your own API keys | ||
| Create | ||
| Documents, PDFs, slides and sheets | ||
| Image generation | Faster[4] | |
| Photo editing | ||
| Video generation | Planned | Planned |
| Music generation | Planned | Planned |
| Research and tools | ||
| Web search | ||
| Deep Research | Up to 200 sources | Up to 1,000 sources |
| App connectors | ||
| AI agents | ||
| Projects | ||
| Knowledge bases | ||
| File library | ||
| Compute and privacy | ||
| Routed to lower-energy compute where available | ||
| Energy summary per account | ||
| Content used to train our models | No by default | No by default |
| History deletion and account export | ||
| Models from Chinese providers | ||
[1] Fair-usage limits apply.
[2] Response times and memory vary by model — per-plan context windows apply where the provider supports them.
[3] Reasoning effort applies to models that support it; other models run at their default setting.
[4] Faster image generation asks the provider to prioritise your request; prioritisation is not guaranteed.
Opal Black provides access to third-party AI models alongside our own model in development. We are not affiliated with, endorsed by, or the owner of third-party models; all trademarks belong to their respective owners.