Models & tiers
Models & Plans
You choose a tier by what you need it to do. Oro picks and routes the underlying model for you. The tier names are deliberately vendor-neutral, so we can swap in a better model without you changing anything.
1. What a tier is
A tier is a level of capability, not a specific model. You pick a tier that fits the work — cheap and fast for high-volume replies, or stronger reasoning for complex conversations — and Oro selects and routes the model that sits behind it.
We keep tier names vendor-neutral on purpose. Because the name describes the job rather than a particular model, we can move a tier to a better model as one becomes available, and nothing changes on your side. You keep the same tier; the work quietly gets better.
2. The classic tiers
Nine tiers, grouped by classification. Each shows what it is best suited to and whether it accepts image and video input.
Lite
Entry-level tiers at the lowest cost.Oro Low
Cheapest tier — fast and economical for high-volume replies.
- Recommended for
- High-volume auto-replies (cheapest)
- Image input
- Yes
- Video input
- Yes
Oro Lite
Vision-capable lite tier — accepts text, image and video at the lowest price point.
- Recommended for
- Image/video-aware replies on a tight budget
- Image input
- Yes
- Video input
- Yes
Standard
Balanced middle tiers for everyday work.Oro Mid
Balanced quality and cost — a solid all-rounder.
- Recommended for
- Balanced replies for most pages
- Image input
- Yes
- Video input
- Yes
Oro One
Fast, low-cost tier with strong multilingual quality including Burmese.
- Recommended for
- Fast multilingual replies
- Image input
- No
- Video input
- No
Recommended
The default tier we suggest for everyday rich-media replies.Oro Flash
Fast vision and video tier with tool calling.
- Recommended for
- Default rich-media replies
- Image input
- Yes
- Video input
- Yes
Pro
Highest reasoning and quality for complex or accuracy-critical work.Oro High
Premium reasoning and tone — for complex or sensitive replies.
- Recommended for
- Complex reasoning, premium tone
- Image input
- Yes
- Video input
- Yes
Oro Two
Advanced reasoning and multi-step planning at a moderate price point.
- Recommended for
- Advanced reasoning, multi-step decisions
- Image input
- No
- Video input
- No
Oro Plus
Stronger reasoning with vision and video — for accuracy-critical media including slip OCR.
- Recommended for
- Slip OCR, fake-detection
- Image input
- Yes
- Video input
- Yes
Oro Hype
Top-tier multi-step reasoning and tool orchestration.
- Recommended for
- Long-horizon content generation
- Image input
- No
- Video input
- No
3. Oro 1.0 cognitive-loop tiers
BetaThe Oro 1.0 tiers run a cognitive loop instead of a single pass. Each turn moves through three stages: a Planner breaks the work into steps, an Executor carries them out, and a Verifier checks the result before it is returned.
When the Verifier is not satisfied, the loop self-corrects and tries again — at most two strikes per turn — so a turn does not run away with itself. These tiers are in beta.
Oro 1.0 Lite
BetaCognitive-loop beta — the fastest pipeline, with Planner and Verifier.
Recommended for cost-sensitive multi-step tasks with verification.
Oro 1.0 Standard
BetaCognitive-loop beta — planning and verification with a fast executor.
Recommended for balanced multi-step quality with reliability gates.
Oro 1.0 Pro
BetaCognitive-loop beta — the strongest planner, executor and verifier.
Recommended for highest-quality multi-step reasoning and execution.
4. Image and video support
Some tiers accept image and video input; others handle text only. This is which is which.
| Tier | Classification | Image input | Video input |
|---|---|---|---|
| Oro Low | Lite | Yes | Yes |
| Oro Lite | Lite | Yes | Yes |
| Oro Mid | Standard | Yes | Yes |
| Oro One | Standard | No | No |
| Oro Flash | Recommended | Yes | Yes |
| Oro High | Pro | Yes | Yes |
| Oro Two | Pro | No | No |
| Oro Plus | Pro | Yes | Yes |
| Oro Hype | Pro | No | No |
| Oro 1.0 Lite (beta) | Lite | No | No |
| Oro 1.0 Standard (beta) | Standard | No | No |
| Oro 1.0 Pro (beta) | Pro | No | No |
5. Bring your own key
If you would rather use your own provider account, you can supply your own Anthropic, OpenAI or Google AI Studio key.
Requests made with your key go straight to that provider, under your own agreement with them, and they consume no Oro credits. The billing and terms for those requests are between you and the provider you brought.
6. Model Context Protocol
Model Context Protocol (MCP) is available on every plan.
7. What each plan costs
Tiers describe what a model does. Plans describe how much you can do each month and what it costs. For the current plans and their prices, see the pricing page.