autofrontierLet the Fleet pick. Balances quality and available capacity for a general request.
- Input
- $6.25/MTok
- Output
- $31.25/MTok
intent1M context128K max outcaller toolsno streaming
CATALOGUE
CompareModel catalogue
These are the only routes Modavio advertises, and each one is priced below at the same constants the billing code uses. Pricing version 2026-08-24.1.
Showing 8 of 8 routes.
auto:cheapfastCheapest qualifying capacity. Capped at the fast tier, so it never bills above fast-tier rates.
- Input
- $1.25/MTok
- Output
- $6.25/MTok
intent200K context64K max outcaller toolsno streaming
auto:codefrontierRouted for code generation, review and refactoring.
- Input
- $6.25/MTok
- Output
- $31.25/MTok
intent1M context128K max outcaller toolsno streaming
auto:reasoningfrontierRouted for analysis and multi-step reasoning.
- Input
- $6.25/MTok
- Output
- $31.25/MTok
intent1M context128K max outcaller toolsno streaming
auto:long-contextfrontierRouted for very large prompts and document work.
- Input
- $6.25/MTok
- Output
- $31.25/MTok
intent1M context128K max outcaller toolsno streaming
haikufastPin the fast tier directly.
- Input
- $1.25/MTok
- Output
- $6.25/MTok
pinned tier200K context64K max outcaller toolsno streaming
sonnetstandardPin the standard tier directly.
- Input
- $3.75/MTok
- Output
- $18.75/MTok
pinned tier1M context128K max outcaller toolsno streaming
opusfrontierPin the frontier tier directly.
- Input
- $6.25/MTok
- Output
- $31.25/MTok
pinned tier1M context128K max outcaller toolsno streaming
FULL SPECIFICATION
| Route | Type | Tier | Context | Max output | Tools | Streaming | Input /MTok | Output /MTok |
|---|---|---|---|---|---|---|---|---|
autoLet the Fleet pick. Balances quality and available capacity for a general request. | Intent | frontier | 1,000K | 128K | Yes | Not yet | $6.25 | $31.25 |
auto:cheapCheapest qualifying capacity. Capped at the fast tier, so it never bills above fast-tier rates. | Intent | fast | 200K | 64K | Yes | Not yet | $1.25 | $6.25 |
auto:codeRouted for code generation, review and refactoring. | Intent | frontier | 1,000K | 128K | Yes | Not yet | $6.25 | $31.25 |
auto:reasoningRouted for analysis and multi-step reasoning. | Intent | frontier | 1,000K | 128K | Yes | Not yet | $6.25 | $31.25 |
auto:long-contextRouted for very large prompts and document work. | Intent | frontier | 1,000K | 128K | Yes | Not yet | $6.25 | $31.25 |
haikuPin the fast tier directly. | Pinned tier | fast | 200K | 64K | Yes | Not yet | $1.25 | $6.25 |
sonnetPin the standard tier directly. | Pinned tier | standard | 1,000K | 128K | Yes | Not yet | $3.75 | $18.75 |
opusPin the frontier tier directly. | Pinned tier | frontier | 1,000K | 128K | Yes | Not yet | $6.25 | $31.25 |
Availability
Every route here is one Modavio can call and price. Modavio does not run a live health probe, so this page states what is offered, not a real-time capacity reading. When the Fleet has no capacity the request fails with upstream_capacity and your reserve is returned in full.
Intents versus pinned tiers
An intent (auto, auto:cheap) lets the Fleet choose the model and reserves at the most expensive tier that intent may resolve to. A pinned tier (haiku, sonnet, opus) reserves and settles at that tier. Either way you are billed for what actually answered.
Tool use
Customer-defined tools ({ name, description, input_schema }) are relayed. Built-in host-capability tools — shell, text editor, computer use — are refused, because they would execute against Modavio infrastructure rather than yours. The Fleet advertises hosted_tools: []. See the docs.
Streaming
Not available yet. Send stream:false or omit it; a streaming request is rejected with a 400 rather than silently buffered, so your client never waits on a stream that will not arrive.