Supported models, API costs & resources
Updated: 11 October 2026
This is the reference page for the model list, API billing and further information mentioned on the Project Wisteria store page.
Using Wisteria
Wisteria is a Windows AI companion that you connect to an AI provider using your own API account and key. The Steam purchase covers the Wisteria application. API usage is billed separately by your selected provider; an AI model or API credit allowance is not included.
Current supported models
- OpenAI GPT-5.6: Sol (gpt-5.6-sol), Terra (gpt-5.6-terra), Luna (gpt-5.6-luna)
- OpenAI GPT-6: Sol (gpt-6-sol), Luna (gpt-6-luna)
- Anthropic Claude: Opus 4.8 (claude-opus-4-8), Sonnet 5 (claude-sonnet-5)
Select the provider and model in Wisteria and enter the corresponding API key.
Approximate API prices
USD per 1 million tokens, checked against the providers' published prices on 11 October 2026. These are standard rates for short-context requests, before taxes and currency conversion.
| Model | Input | Cached input | Output |
|---|---|---|---|
| GPT-5.6 Sol | $4.00 | $0.40 | $20.00 |
| GPT-5.6 Terra | $2.00 | $0.20 | $12.00 |
| GPT-5.6 Luna | $0.20 | $0.02 | $1.20 |
| GPT-6 Sol | $2.00 | $0.20 | $10.00 |
| GPT-6 Luna | $0.10 | $0.01 | $0.50 |
| Claude Opus 4.8 | $5.00 | $0.50 | $25.00 |
| Claude Sonnet 5 | $2.00 | $0.20 | $10.00 |
For a simple example, one request with 10,000 uncached input tokens and 1,000 output tokens costs about $0.0015 with GPT-6 Luna, $0.03 with GPT-6 Sol or Claude Sonnet 5, and $0.075 with Claude Opus 4.8. These examples show the arithmetic, rather than a fixed price per message.
Input includes the conversation and memory context prepared for the request. Images, memory compression and other model requests also contribute to usage. The actual total depends on the selected model, tokens and cache usage. Creating a cache may be charged separately: OpenAI's listed cache-write rates are 1.25 times input; Claude's 5-minute and 1-hour cache writes are 1.25 and 2 times input. Longer-context, regional or service-tier pricing can differ. Wisteria's cost display is an estimate; the provider's usage dashboard and invoice are the billing reference.
Approximate costs observed in actual use
The August–September 2026 usage data described on our official website showed average input costs of about $0.08–0.15 per request for flagship models and $0.02–0.05 for mid-tier models. With long-term memory covering about 7,000 turns, approximately 60–80% of input tokens were reused from cache. These historical observations exclude output costs and are separate from the token arithmetic examples above.
Official site and detailed references
- Wisteria official website — product, memory structure and further resources
- Wisteria official website in Korean
- Developer contact and support
- OpenAI API pricing
- Anthropic Claude API pricing
The model-specific OpenAI price pages used for this table are GPT-5.6 Sol, GPT-5.6 Terra, GPT-5.6 Luna, GPT-6 Sol and GPT-6 Luna. Model availability and provider prices may change; this announcement is the central reference for future updates.