SillyTavern presets. Import one, then mind the Context Size.

A Chat Completion preset holds your prompts and sampler settings, and its Context Size decides how much of the chat goes with every message. That is your bill: DeepSeek V4.1 Flash gives about 7,900 messages for $10 at 8,192 tokens and about 1,100 at 65,536.

Import a preset

  • Open AI Response Configuration: Click the sliders icon in the top bar. Chat Completion Presets sits at the top of the panel.
  • Import preset: Click Import preset and pick the preset file, a JSON. It becomes the selected preset.
  • Save your changes: After changing a setting, click Update current preset, or Export preset to share it.
  • Keep one per connection: Bind presets to API connections, if you switch between models with different needs.

The settings that matter

  • Context Size (tokens): How much of the chat goes with every message. More context remembers more and costs more. The slider stops at 4,095 until you tick Unlocked Context Size.
  • Max Response Length (tokens): The longest reply the model may write. Replies are billed at the output price, the higher of the two.
  • Temperature: How varied the replies are. Lower is steadier, higher is more surprising; it does not change the price.

What $10 buys at each Context Size

Once a chat fills the context, every message sends about that many tokens. Each reply here is 400 tokens, priced without caching.

  • 8,192 tokens: DeepSeek V4.1 Flash 7,900, DeepSeek V4 Pro 1,900, Gemini 3.8 Flash 1,600, Claude Sonnet 4.6 400, Claude Opus 5.5 320
  • 16,384 tokens: DeepSeek V4.1 Flash 4,200, DeepSeek V4 Pro 1,000, Gemini 3.8 Flash 900, Claude Sonnet 4.6 220, Claude Opus 5.5 180
  • 32,768 tokens: DeepSeek V4.1 Flash 2,200, DeepSeek V4 Pro 530, Gemini 3.8 Flash 470, Claude Sonnet 4.6 110, Claude Opus 5.5 90
  • 65,536 tokens: DeepSeek V4.1 Flash 1,100, DeepSeek V4 Pro 270, Gemini 3.8 Flash 240, Claude Sonnet 4.6 60, Claude Opus 5.5 40

If a preset misbehaves

  • Stuck at 4,095: Tick Unlocked Context Size above the slider, then set the size you want.
  • Import shows nothing: The file is a preset for another API, such as Text Completion. Chat Completion presets only show when the API is Chat Completion.
  • Replies cut off: Raise Max Response Length (tokens) in the same panel, then Update current preset.

Connect a model

A preset needs a model to talk to. Paste these into API Connections (the plug icon); for the key, start on the $2 trial credit, or top up from $10 below and it arrives by email.

API: Chat Completion
Chat Completion Source: Custom (OpenAI-compatible)
Custom Endpoint (Base URL): https://api.apisrouter.com/v1
Custom API Key: sk-your-apisrouter-key
Enter a Model ID: deepseek-v4.1-flash

Two ways to pay

Both fill one balance, spent as you chat.

FAQ

What is a SillyTavern preset?

A saved set of Chat Completion settings: the prompts SillyTavern sends and sampler settings such as Temperature, Context Size and Max Response Length. Presets are shared as JSON files.

How do I import a preset in SillyTavern?

Open AI Response Configuration (the sliders icon), then click Import preset under Chat Completion Presets and pick the JSON file.

Why can I not set Context Size above 4,095?

The slider is locked until you tick Unlocked Context Size, above it in the same panel.

What Context Size should I use?

As much as the chat needs to stay coherent, and no more: every token of context goes with every message. On DeepSeek V4.1 Flash, $10 buys about 7,900 messages at 8,192 tokens and about 1,100 at 65,536.

Does a preset work with every model?

A preset works with any Chat Completion source, including Custom (OpenAI-compatible) with APIsRouter. The same key runs DeepSeek V4.1 Flash, Claude Sonnet 4.6 and Claude Opus 5.5.

Can I try it for free?

Yes. Sign in and claim the $2 trial credit from your dashboard, no card needed: about 1,600 DeepSeek V4.1 Flash messages of 8,000 tokens each.