Pricing & Credits
ORQO charges through a three-part model: a monthly subscription for platform access, credits for each LLM call your agents make, and a Heavy Lift surcharge on calls with oversized context. You can switch to your own LLM API keys at any time to cut per-call credit costs dramatically.
1 credit = $0.01. Credits are the unit of consumption across the platform — agent calls, heavy-context surcharges, per-use marketplace apps.
Plans
| Starter | Pro | Growth | Enterprise | |
|---|---|---|---|---|
| Price | Free | $49/mo | $149/mo | $499/mo |
| Annual | — | $39/mo (save 20%) | $119/mo (save 20%) | $399/mo (save 20%) |
| Credits included | 100/mo | 3,000/mo | 12,000/mo | 50,000/mo |
| Seats | 2 | 10 | 25 | Unlimited |
| Projects | 1 | 5 | 20 | Unlimited |
| Teams | 1 | 5 | 15 | Unlimited |
| Agents per team | 3 | 10 | 25 | Unlimited |
| Workflows | 2 | 15 | 50 | Unlimited |
| Runs per month | 20 | 300 | 1,500 | Unlimited |
| Knowledge Graph | Read-only | Full | Full | Full + API |
| Channels | 1 | 3 | All | All + custom |
| Marketplace apps | 2 | 10 | Unlimited | Unlimited + private |
| Support | Community | Priority | Dedicated |
All plans include the visual workflow and team builders, the knowledge graph, marketplace access, and multi-agent orchestration. No credit card required for Starter.
How Credits Work
Every LLM call an agent makes during a workflow run costs credits. The amount depends on which model is used.
A typical run uses around a dozen LLM calls. Depending on your model mix, that's roughly 12–56 credits per run.
Agent Action Units (AAUs)
Raw LLM prices fluctuate daily and vary wildly — some models cost $0.02 per million input tokens, others cost $30. Pricing your workflows directly against that moving target would be unpredictable for you and impossible for us to margin. So we abstracted it.
An Agent Action Unit (AAU) is one agentic turn — one LLM call made by one agent. The credit cost per AAU depends on the model's tier, determined by its input price per million tokens.
The Five Tiers
Every model in ORQO carries a colored dot so you can see its cost at a glance — in the model library, in the team builder, and in the consumption dashboard.
| Dot | Tier | Credits | Input price (inclusive) | Example models |
|---|---|---|---|---|
| 🟦 | Efficient | 1 | ≤ $0.15 / M | GPT-5-nano, Gemini Flash Lite, Ministral |
| 🟩 | Standard | 3 | $0.15 – $1.00 / M | Gemini 2.5 Flash, GPT-4.1-mini, MiMo-V2-Pro, Grok-Fast, DeepSeek-R1 |
| 🟨 | Performance | 10 | $1.00 – $5.00 / M | Claude Sonnet, GPT-4.1, Gemini Pro, Haiku 4.5, o3-mini, o4-mini |
| 🟧 | Premium | 25 | $5.00 – $15.00 / M | Claude Opus 4.5, o3-deep-research |
| 🟥 | Frontier | 60 | > $15.00 / M | Claude Opus 4, o1, o3-pro, GPT-5-pro |
Upper bounds are inclusive — a model priced exactly at the cutoff sits in that tier, not the one above.
Use 🟦 Efficient models for classification, summarization, and routing. Reach for 🟨 Performance when tasks need deeper reasoning. A workflow that spends 🟦 on 8 calls and 🟨 on 4 often costs less than one that uses 🟩 Standard for all 12.
What Can I Run on Each Plan?
| Run type | Credits per run | Starter | Pro | Growth |
|---|---|---|---|---|
| Budget (mostly 🟦🟩) | ~20 | 5 runs | 150 runs | 600 runs |
| Typical (mixed tiers) | ~38 | 2 runs | 79 runs | 316 runs |
| Performance (mostly 🟨) | 56 | 1 run | 53 runs | 214 runs |
| Premium (🟧 mix) | 152 | — | 19 runs | 78 runs |
These are approximations for workflows at the production average of ~12 calls per run. Long or document-heavy workflows cost proportionally more.
Heavy Lift
LLM calls with very large context windows cost providers more — and they cost ORQO more to orchestrate (classification, compaction, analysis all scale with context length). A small Heavy Lift surcharge applies when a single call exceeds 30,000 tokens:
| Total tokens in the call | Extra credits |
|---|---|
| ≤ 30,000 | No surcharge |
| 30,001 – 60,000 | +1 credit per 10K tokens above 30K |
| 60,001+ | +3 credits per 10K tokens above 60K |
Most calls sit well under the threshold. Heavy Lift is designed to catch document-heavy workflows where a single agent reads a 50-page PDF in one turn — not a tax on every call.
The 30K threshold applies after prompt caching — if a provider returns a cached prefix, only the freshly processed tokens count toward Heavy Lift. Calls that run entirely from cache pay no surcharge.
Heavy Lift applies equally to BYOT calls: even when you pay the LLM provider directly, heavy-context calls cost ORQO more to orchestrate.
Bring Your Own Key
Already have API keys from OpenAI, Anthropic, Google, or OpenRouter? Add them as credentials and your workflows will use them instead of ORQO's platform keys. You pay the LLM provider directly for token usage. ORQO credits cover only the orchestration overhead.
BYOT Credit Rates
| Calls per run | BYOT credit rate |
|---|---|
| Calls 1 – 15 | 1 credit each |
| Calls 16+ | 2 credits each |
Compare that against the platform-key tier rates:
| Mode | 🟦 | 🟩 | 🟨 | 🟧 | 🟥 |
|---|---|---|---|---|---|
| Platform keys | 1 | 3 | 10 | 25 | 60 |
| Your keys | 1–2 | 1–2 | 1–2 | 1–2 | 1–2 |
A Growth subscriber running on 🟨 Performance models pays 10 credits per call on platform keys. On their own key, the same call costs 1 credit — a 10× reduction in credit burn. You still pay the LLM provider, but you cut ORQO credit consumption by the same tier multiplier.
Mix and Match
You can bring keys for some providers and use ORQO's for others. Resolution happens per call — each call checks whether you've stored a credential for that provider:
- You have an OpenAI credential → OpenAI calls use your key (BYOT rate)
- You have no Anthropic credential → Anthropic calls use ORQO's key (tier rate)
No configuration switches, no manual routing. Just add a credential when you're ready.
Why BYOT?
- Lower credit burn. The same workflow costs 5–10× fewer credits.
- Compliance. Enterprise accounts often require their own API contracts, audit trails, and data-residency controls.
- Provider diversity. Bring keys for providers ORQO doesn't have platform tokens for.
- Stability. Your rate is tied to your provider contract, not our cost structure.
Top-Ups
Need more credits before the month ends? Add them anytime.
| Plan | Top-up rate |
|---|---|
| Starter | $0.012 per credit |
| Pro | $0.011 per credit |
| Growth | $0.010 per credit |
| Enterprise | Custom |
Minimum top-up: $1.00 (100 credits). Purchased credits never expire — they stay on your account until used.
Credit Rollover
Included monthly credits reset each billing cycle. Unused credits don't carry forward — monthly credits are use-it-or-lose-it, standard SaaS practice.
Purchased credits (top-ups) never expire. Once you've paid for them, they're yours.
Consumption order: Included credits are consumed first, then purchased credits. That way you always burn the expiring bucket before touching the permanent one.
Changing Plans
Upgrades take effect immediately — you're billed the pro-rated difference and get access to the new plan's credit allowance on the spot.
Downgrades take effect at the end of your current billing period. You keep access to your current plan's features until then.
FAQ
How do I know how much a run will cost before I start it? Each model carries its tier dot in the team builder, model library, and consumption dashboard. Multiply the agent's model tier by the expected number of calls to estimate. A dedicated pre-run estimator is on the roadmap.
What happens when I run out of credits? You can still access everything — edit workflows, view past runs, browse your knowledge graph, configure teams. You just can't start new workflow runs until you top up or your monthly credits renew.
Can the tier of a model change? Yes — if a provider changes their pricing, a model may move tiers. We never raise a model's tier without advance notice via changelog and in-app banner. Models going down a tier (cheaper for you) happen silently.
Do marketplace apps cost extra? Some apps are free. Others charge per use (a credit deduction) or a monthly subscription (separate Stripe charge) — set by the app developer. ORQO takes a 20% platform fee from developer revenue. Any LLM calls the app triggers during your workflows consume your credits at the standard tier rates.
What about reasoning models that think before answering? Reasoning tokens (o1, o3, o4-mini, Claude extended thinking) count as output tokens in the provider's billing, and the tier dot reflects that. A model that charges $3/M for output but $15/M for reasoning still gets classified by its input price. The Heavy Lift surcharge catches runaway thinking calls by counting total tokens in the call, including reasoning.
Why credits instead of just passing through provider costs? Raw provider prices change frequently and vary 1000× across the model landscape. Pricing in credits lets you plan: one call on 🟨 Performance always costs 10 credits, whether the provider's price moves up or down. It also lets us maintain a consistent margin that funds the platform you're using — the visual builders, the knowledge graph, the multi-channel integrations, the security layer.
I'm on Starter and keep running out. What should I do? Either top up ($1 = 100 credits), add your own OpenRouter key to switch to BYOT rates, or upgrade to Pro. Most solo users find Pro covers 1–3 active workflows comfortably.