> ## Documentation Index
> Fetch the complete documentation index at: https://docs.capy.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Models

> Every model Capy serves and what it costs, plus how routes, subscriptions, fast mode, reasoning effort, and BYOK actually behave.

Capy's model picker lists entries, and an entry is a route (the model plus who serves and bills it): the same weights served through Capy's API key and through your Codex or Copilot subscription are two different entries with two different billing stories. `openai/gpt-5.6-sol` bills Capy credits at the rates below; `codex/gpt-5.6-sol` runs on your ChatGPT subscription. The picker labels each entry's billing story, and picking one picks both the model and who pays.

## Your selection is a lock

Your selection pins the entry, and there is never cross-entry fallback. If your Codex subscription hits its weekly limit or you disconnect it, the run fails visibly, naming the lost connection; Capy never silently reroutes onto its own billing or another provider. That's deliberate: a silent reroute is a surprise bill, while a visible failure is one click to fix (reconnect, or pick another entry).

The one exception is retirement, and it can't change your deal: when a model is retired, Capy redirects to the entry's designated replacement mid-thread at the next step, with a visible notice naming both models. The replacement is guaranteed to sit in the same billing class, so a retired Capy-billed model never lands you on a subscription route or vice versa.

## Capy-billed models

Prices are per 1M tokens on the Capy-billed route. Models marked BYOK can also run on your own provider API key from [Settings → Models](https://capy.ai/settings/models).

| Model                      | ID                                | Context | Input (per 1M)                                                    | Output (per 1M)                      | BYOK |
| -------------------------- | --------------------------------- | ------- | ----------------------------------------------------------------- | ------------------------------------ | ---- |
| **GPT-5.6 Sol** (default)  | `openai/gpt-5.6-sol`              | 1.05M   | ≤272K: \$5.00 (cached \$0.50)<br />>272K: \$10.00 (cached \$1.00) | ≤272K: \$30.00<br />>272K: \$45.00   | ✓    |
| **GPT-5.6 Terra**          | `openai/gpt-5.6-terra`            | 1.05M   | ≤272K: \$2.50 (cached \$0.25)<br />>272K: \$5.00 (cached \$0.50)  | ≤272K: \$15.00<br />>272K: \$22.50   | ✓    |
| **GPT-5.6 Luna**           | `openai/gpt-5.6-luna`             | 1.05M   | ≤272K: \$1.00 (cached \$0.10)<br />>272K: \$2.00 (cached \$0.20)  | ≤272K: \$6.00<br />>272K: \$9.00     | ✓    |
| **GPT-5.5**                | `openai/gpt-5.5`                  | 1.05M   | ≤272K: \$5.00 (cached \$0.50)<br />>272K: \$10.00 (cached \$1.00) | ≤272K: \$30.00<br />>272K: \$45.00   | ✓    |
| **GPT-5.5 Pro**            | `openai/gpt-5.5-pro`              | 1.05M   | ≤272K: \$30.00<br />>272K: \$60.00                                | ≤272K: \$180.00<br />>272K: \$270.00 | ✓    |
| **GPT-5.4 Mini**           | `openai/gpt-5.4-mini`             | 400K    | \$0.75 (cached \$0.075)                                           | \$4.50                               | ✓    |
| **GPT-5.4**                | `openai/gpt-5.4`                  | 1.05M   | ≤272K: \$2.50 (cached \$0.25)<br />>272K: \$5.00 (cached \$0.50)  | ≤272K: \$15.00<br />>272K: \$22.50   | ✓    |
| **GPT-5.3 Codex**          | `openai/gpt-5.3-codex`            | 400K    | \$1.75 (cached \$0.175)                                           | \$14.00                              | ✓    |
| **Claude Fable 5**         | `anthropic/claude-fable-5`        | 1M      | \$10.00 (cached \$1.00)                                           | \$50.00                              | ✓    |
| **Claude Opus 5**          | `anthropic/claude-opus-5`         | 1M      | \$5.00 (cached \$0.50)                                            | \$25.00                              | ✓    |
| **Claude Sonnet 5**        | `anthropic/claude-sonnet-5`       | 1M      | \$2.00 (cached \$0.20)                                            | \$10.00                              | ✓    |
| **Claude Haiku 4.5**       | `anthropic/claude-haiku-4-5`      | 200K    | \$1.00 (cached \$0.10)                                            | \$5.00                               | ✓    |
| **Claude Opus 4.8**        | `anthropic/claude-opus-4-8`       | 1M      | \$5.00 (cached \$0.50)                                            | \$25.00                              | ✓    |
| **Claude Opus 4.7**        | `anthropic/claude-opus-4-7`       | 1M      | \$5.00 (cached \$0.50)                                            | \$25.00                              | ✓    |
| **Claude Sonnet 4.6**      | `anthropic/claude-sonnet-4-6`     | 1M      | \$3.00 (cached \$0.30)                                            | \$15.00                              | ✓    |
| **Claude Opus 4.6**        | `anthropic/claude-opus-4-6`       | 1M      | \$5.00 (cached \$0.50)                                            | \$25.00                              | ✓    |
| **Claude Opus 4.5**        | `anthropic/claude-opus-4-5`       | 200K    | \$5.00 (cached \$0.50)                                            | \$25.00                              | ✓    |
| **Grok 4.5**               | `xai/grok-4.5`                    | 500K    | ≤200K: \$2.00 (cached \$0.30)<br />>200K: \$4.00 (cached \$0.60)  | ≤200K: \$6.00<br />>200K: \$12.00    | ✓    |
| **Grok 4.3**               | `xai/grok-4.3`                    | 1M      | ≤200K: \$1.25 (cached \$0.20)<br />>200K: \$2.50 (cached \$0.40)  | ≤200K: \$2.50<br />>200K: \$5.00     | ✓    |
| **Kimi K3**                | `moonshotai/kimi-k3`              | 1.05M   | \$3.00 (cached \$0.30)                                            | \$15.00                              |      |
| **Kimi K2.7 Code**         | `moonshotai/kimi-k2.7-code`       | 262K    | \$0.95 (cached \$0.19)                                            | \$4.00                               |      |
| **Kimi K2.6**              | `moonshotai/kimi-k2.6`            | 262K    | \$0.95 (cached \$0.16)                                            | \$4.00                               |      |
| **GLM-5.2**                | `zai/glm-5.2`                     | 1.05M   | \$1.40 (cached \$0.26)                                            | \$4.40                               |      |
| **GLM-5.1**                | `zai/glm-5.1`                     | 203K    | \$1.40 (cached \$0.26)                                            | \$4.40                               |      |
| **GLM-5V-Turbo**           | `zai/glm-5v-turbo`                | 203K    | \$1.20 (cached \$0.24)                                            | \$4.00                               |      |
| **GLM-5-Turbo**            | `zai/glm-5-turbo`                 | 203K    | \$1.20 (cached \$0.24)                                            | \$4.00                               |      |
| **DeepSeek V4 Pro**        | `deepseek/deepseek-v4-pro`        | 1.05M   | \$1.74 (cached \$0.145)                                           | \$3.48                               |      |
| **DeepSeek V4 Flash 0731** | `deepseek/deepseek-v4-flash-0731` | 1.05M   | \$0.14 (cached \$0.028)                                           | \$0.28                               |      |
| **Gemini 3.1 Pro Preview** | `google/gemini-3.1-pro-preview`   | 1.05M   | ≤200K: \$2.00 (cached \$0.20)<br />>200K: \$4.00 (cached \$0.40)  | ≤200K: \$12.00<br />>200K: \$18.00   | ✓    |
| **Gemini 3 Flash Preview** | `google/gemini-3-flash-preview`   | 1.05M   | \$0.50 (cached \$0.05)                                            | \$3.00                               | ✓    |
| **Qwen3.8 Max**            | `qwen/qwen3.8-max`                | 1M      | \$2.00 (cached \$0.25)                                            | \$6.00                               |      |

## Subscription and organization routes

These entries run through a connected provider subscription (Codex, Copilot, SuperGrok) or an organization-connected account (Azure), and model usage bills to that subscription or account instead of Capy credits. They appear in the picker only while the connection exists, and a subscription route may serve a smaller context window than the Capy entry for the same model: Codex caps at 400K where `openai/gpt-5.6-sol` serves 1.05M.

| Model                      | ID                               | Runs through         | Context |
| -------------------------- | -------------------------------- | -------------------- | ------- |
| **GPT-5.6 Sol**            | `codex/gpt-5.6-sol`              | Codex                | 400K    |
| **GPT-5.6 Terra**          | `codex/gpt-5.6-terra`            | Codex                | 400K    |
| **GPT-5.6 Luna**           | `codex/gpt-5.6-luna`             | Codex                | 400K    |
| **GPT-5.5**                | `codex/gpt-5.5`                  | Codex                | 400K    |
| **GPT-5.4 Mini**           | `codex/gpt-5.4-mini`             | Codex                | 400K    |
| **GPT-5.4**                | `codex/gpt-5.4`                  | Codex                | 400K    |
| **GPT-5.3 Codex Spark**    | `codex/gpt-5.3-codex-spark`      | Codex                | 128K    |
| **GPT-5.6 Sol**            | `copilot/gpt-5.6-sol`            | Copilot              | 1.05M   |
| **GPT-5.6 Terra**          | `copilot/gpt-5.6-terra`          | Copilot              | 1.05M   |
| **GPT-5.6 Luna**           | `copilot/gpt-5.6-luna`           | Copilot              | 1.05M   |
| **Claude Opus 4.8**        | `copilot/claude-opus-4-8`        | Copilot              | 200K    |
| **GPT-5.5**                | `copilot/gpt-5.5`                | Copilot              | 1.05M   |
| **Claude Opus 4.7**        | `copilot/claude-opus-4-7`        | Copilot              | 200K    |
| **GPT-5.4 Mini**           | `copilot/gpt-5.4-mini`           | Copilot              | 400K    |
| **GPT-5.4**                | `copilot/gpt-5.4`                | Copilot              | 1.05M   |
| **Gemini 3.1 Pro Preview** | `copilot/gemini-3.1-pro-preview` | Copilot              | 1M      |
| **Claude Sonnet 4.6**      | `copilot/claude-sonnet-4-6`      | Copilot              | 200K    |
| **GPT-5.3 Codex**          | `copilot/gpt-5.3-codex`          | Copilot              | 400K    |
| **Claude Opus 4.6**        | `copilot/claude-opus-4-6`        | Copilot              | 200K    |
| **Gemini 3 Flash Preview** | `copilot/gemini-3-flash-preview` | Copilot              | 128K    |
| **Claude Opus 4.5**        | `copilot/claude-opus-4-5`        | Copilot              | 200K    |
| **Claude Haiku 4.5**       | `copilot/claude-haiku-4-5`       | Copilot              | 200K    |
| **Grok 4.5**               | `supergrok/grok-4.5`             | SuperGrok            | 500K    |
| **Composer 2.5 Fast**      | `supergrok/composer-2.5-fast`    | SuperGrok            | 200K    |
| **GPT-5.6 Sol**            | `azure/gpt-5.6-sol`              | Azure (organization) | 1.05M   |
| **Claude Fable 5**         | `azure/claude-fable-5`           | Azure (organization) | 1M      |

## Fast mode

Entries with a Fast toggle in the selector run through the provider's priority tier at a higher token rate. When a Capy entry is served by several internal hosts, fast filters to the hosts that offer it; when none is available, the request serves at the standard tier and you're billed for what was actually served, never fast rates for a standard response. Fast rates per 1M tokens on the Capy-billed route:

| Model               | Fast input (per 1M)     | Fast output (per 1M) |
| ------------------- | ----------------------- | -------------------- |
| **GPT-5.6 Sol**     | \$10.00 (cached \$1.00) | \$60.00              |
| **GPT-5.6 Terra**   | \$5.00 (cached \$0.50)  | \$30.00              |
| **GPT-5.6 Luna**    | \$2.00 (cached \$0.20)  | \$12.00              |
| **GPT-5.5**         | \$12.50 (cached \$1.25) | \$75.00              |
| **GPT-5.4 Mini**    | \$1.50 (cached \$0.15)  | \$9.00               |
| **GPT-5.4**         | \$5.00 (cached \$0.50)  | \$30.00              |
| **GPT-5.3 Codex**   | \$3.50 (cached \$0.35)  | \$28.00              |
| **Claude Opus 5**   | \$10.00 (cached \$1.00) | \$50.00              |
| **Claude Opus 4.8** | \$10.00 (cached \$1.00) | \$50.00              |

Pro mode, on the entries that offer it, bills at standard token rates and, unlike fast, never degrades: a pro request either serves pro or fails.

## Reasoning effort

Most models expose a reasoning effort, from `none` up to `max`; each entry supports its own subset, shown in the picker. An effort the selected entry doesn't support fails before the request is sent; Capy never quietly approximates it with a different setting.

## Naming a model in a prompt

You don't have to type entry ids. Say `sol`, `opus`, or `gpt-5.6-sol` in a message and Capy resolves the phrase to exactly one connected route, or asks; it never guesses. Matching is whole-string, so `5.6-sol` matches nothing: a fragment that resolved would be a typo silently picking a model. When a phrase maps to several connected routes (`sol` on both your Codex subscription and Capy billing), Capy asks once and remembers the answer for your whole organization, so teammates don't answer the same question differently. An unknown phrase is refused with up to three ranked suggestions, and a phrase that names only unconnected routes is refused by naming the missing connection.

## Bring your own key

BYOK is a payment toggle on a Capy entry, never a separate picker item. Activate it in [Settings → Models](https://capy.ai/settings/models) with your provider API key: routing pins to the hosts your key can serve (first-party, not Bedrock or Vertex), model usage bills to your provider account, and the picker relabels the entry as running on your key. The entry, its modes, and its facts are otherwise identical.

## Per-task model choice

Threads and their tasks pick models independently: when the agent spawns a task, it can run that task on a different entry: a cheaper model for mechanical work, a stronger one for a hard subsystem. A task that names no model runs on the thread's own.