Skip to main content

Degradation Model: Tiers and Limits

The Degradation Model defines how Zeq capabilities scale across subscription tiers. Free-tier users receive a degraded experience (lower precision, fewer operators, rate limits), while paying users unlock monotonically more capabilities—until Architect tier, which has no limits.

info

"Degradation" means that free-tier capabilities are a strict subset of paid-tier capabilities. Each tier strictly improves upon the previous one.

The Four Paid Tiers (+ Free)

TierPriceRate LimitPrecisionOperatorsAdvanced ProtocolsUse Case
Free$0/mo10 req/min≤1% error42 core onlyLockedLearning, prototyping
Starter$29/mo100 req/min≤0.5% error234 allPartial (70%)Small projects, research
Builder$99/mo500 req/min≤0.1% error234 allFull (100%)Medium projects, production
Advanced$499/mo5000 req/min≤0.01% error234 allFull (100%)Complex simulations, teams
Architect$2499/moUnlimited≤0.001% error234 allFull (100%)Enterprise, research orgs

Free Tier: The Degraded Experience

Free users can:

  • Execute 10 computations per minute (hard limit; requests beyond this are rejected with 429 Too Many Requests)
  • Use only 42 core operators (out of 1,536 available)
  • Access only basic operators: QM1, NM19, GR1, CS1, EM1, SM1, and others (one per major domain)
  • Receive results with ≤1% precision error (relaxed tolerance)
  • Cannot access advanced protocols (see below)
  • Results are marked as "Free Tier" in metadata (not suitable for publication)

Free Tier Rate Limiting

curl -X POST https://zeq.dev/api/zeq/compute \
-H "Authorization: Bearer YOUR_API_KEY"

If you exceed 10 computations per minute:

Response (429 Too Many Requests):

{
"error": "RATE_LIMIT_EXCEEDED",
"tier": "free",
"limit": {
"computations_per_minute": 10,
"window_seconds": 60
},
"current_usage": {
"computations_this_minute": 11,
"quota_exceeded_by": 1
},
"retry_after_seconds": 45,
"upgrade_message": "Upgrade to Starter tier for 100 req/min, or Builder for 500 req/min."
}

Free Tier Operator List (42 core operators)

QM1   (Schrödinger)
NM19 (Newton)
GR1 (Einstein Field Equations - basic)
CS1 (Complexity – basic algorithm analysis)
EM1 (Maxwell Equations – basic)
SM1 (Boltzmann Distribution)
OD1 (Linear ODE – basic)
PD1 (Laplace Equation – basic)
PR1 (Probability – basic)
ST1 (Statistics – descriptive)
AC1 (Acoustics – basic)
OP1 (Optics – basic)
TH1 (Thermodynamics – basic)
FL1 (Fluid Dynamics – basic inviscid)
... and 28 more basic operators (one per domain)

Accessing any operator outside this list results in an error:

curl -X POST https://zeq.dev/api/zeq/compute \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"operators": ["QM23"]
}'

Response (403 Forbidden):

{
"error": "OPERATOR_NOT_ACCESSIBLE",
"tier": "free",
"requested_operator": "QM23",
"reason": "QM23 (advanced quantum mechanics) is available only on Starter tier and above.",
"accessible_in_tier": "free",
"operators_in_free_tier": 42,
"upgrade_suggestion": "Upgrade to Starter tier to access 234 operators."
}

Free Tier Precision

Free-tier results have a relaxed precision guarantee of ≤1% error. For scientific work, this is often insufficient.

Example: A free-tier QM1 computation might return:

{
"result": 4.852103847,
"precision_error_percent": 0.87,
"tier": "free",
"note": "Free tier results are for educational purposes only. Publication-quality results require paid tiers."
}

Starter Tier: Entry-Level Production

Starter users get:

  • 100 computations per minute (10× free)
  • All 234 operators
  • ≤0.5% precision error (good for most research)
  • 70% of advanced protocols (see below)
  • Results are suitable for academic publication (with proper attribution)

Price: $29/month

When to Use Starter

  • Running research projects with moderate computational load
  • Publishing in peer-reviewed journals (results verified for accuracy)
  • Teaching and coursework (students working on real science)
  • Exploring all operators without restriction

Starter Tier Example

curl -X POST https://zeq.dev/api/zeq/compute \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"operators": ["QM23", "EM12", "NM19"],
"parameters": {...}
}'

Response (Success):

{
"result": 4.852103847,
"precision_error_percent": 0.34,
"tier": "starter",
"operators_used": ["QM23", "EM12", "NM19", "KO42"],
"zeqproof": {...},
"advanced_protocols_available": "70%"
}

Builder Tier: Production Standard

Builder users get:

  • 500 computations per minute (50× free)
  • All 234 operators
  • ≤0.1% precision error (tight, suitable for publication and regulatory work)
  • 100% of advanced protocols (unrestricted access)
  • Results are suitable for regulatory submissions (FDA, EMA, etc.)

Price: $99/month

When to Use Builder

  • Production-grade applications (fintech, pharma, engineering)
  • Regulatory submissions (FDA drug trials, safety-critical systems)
  • Academic research with high precision requirements
  • Teams working on shared projects

Advanced Tier: Research + Teams

Advanced users get:

  • 5000 computations per minute (500× free)
  • All 234 operators
  • ≤0.01% precision error (extremely tight)
  • 100% of advanced protocols
  • Dedicated API endpoint (lower latency)
  • Team collaboration (multiple users on one account)
  • Priority support (24/7 SLA)

Price: $499/month

When to Use Advanced

  • Large-scale simulations (climate modeling, structural analysis)
  • Teams with multiple users (research labs, companies)
  • Real-time applications (fintech, robotics)
  • Complex scientific projects requiring tight coordination

Architect Tier: Unlimited

Architect users get:

  • Unlimited computations (no rate limit)
  • All 234 operators
  • ≤0.001% precision error (maximum accuracy)
  • 100% of advanced protocols
  • Custom infrastructure (on-premises or dedicated cloud)
  • Dedicated account manager
  • Custom SLAs

Price: $2499/month (or custom enterprise agreement)

When to Use Architect

  • Enterprise organizations (Fortune 500, national labs)
  • Mission-critical applications (space systems, nuclear, medical devices)
  • Research organizations publishing dozens of papers per year
  • Custom requirements (on-premises deployment, audit trails, etc.)

Precision Tiers: What They Mean

TierPrecisionError BoundSuitable For
Free (≤1%)Worst± 0.01 (for result ~1.0)Learning only
Starter (≤0.5%)Good± 0.005 (for result ~1.0)Academic publication
Builder (≤0.1%)Excellent± 0.001 (for result ~1.0)Regulatory work
Advanced (≤0.01%)Very High± 0.0001 (for result ~1.0)Engineering, fintech
Architect (≤0.001%)Maximum± 0.00001 (for result ~1.0)Critical systems

Note: These are relative errors. For an absolute result of 100.0:

  • Free tier: ± 1.0
  • Builder tier: ± 0.1
  • Architect tier: ± 0.01

Advanced Protocols

Advanced protocols are domain-specific, specialized computational methods that unlock additional capabilities:

  • Adaptive Mesh Refinement (AMR): Automatically refines spatial grid in high-gradient regions
  • Parallel Time Integration: Solves for multiple time steps in parallel (speeds up long simulations)
  • Quantum Error Correction: Reduces errors in quantum computations
  • Multi-Scale Coupling: Simultaneously solves problems at different length/timescales
  • Uncertainty Quantification: Quantifies sensitivity to parameter variations
  • Machine Learning Emulation: Learns fast surrogate models for expensive computations
  • ... and 227 more advanced protocols

Availability by Tier:

TierAccessExample Restriction
Free0/234All locked
Starter164/234AMR, Parallel Time, ML Emulation locked
Builder234/234All unlocked
Advanced234/234All unlocked
Architect234/234All unlocked

Using an Unavailable Protocol (Free Tier)

curl -X POST https://zeq.dev/api/zeq/compute \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"operators": ["QM1"],
"advanced_protocol": "adaptive_mesh_refinement"
}'

Response (403 Forbidden):

{
"error": "ADVANCED_PROTOCOL_NOT_AVAILABLE",
"tier": "free",
"requested_protocol": "adaptive_mesh_refinement",
"reason": "Advanced protocols are available only on Starter tier (70% access) and above.",
"upgrade_paths": [
{
"tier": "starter",
"price_monthly": 29,
"protocols_unlocked": 164
},
{
"tier": "builder",
"price_monthly": 99,
"protocols_unlocked": 234
}
]
}

Credits and Overage Pricing

For paid tiers, you can exceed your monthly rate limit by purchasing overages:

curl -X POST https://zeq.dev/api/zeq/purchase-credits \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"tier": "starter",
"credit_amount": 1000,
"auto_refill": true,
"refill_threshold": 100,
"refill_amount": 500
}'

Pricing:

  • 1 credit = 1 computation beyond your monthly allowance
  • Free tier: Cannot purchase credits (upgrade required)
  • Starter: $0.03 per 100 extra computations
  • Builder: $0.02 per 100 extra computations
  • Advanced: $0.01 per 100 extra computations
  • Architect: Custom pricing

Response:

{
"purchase_confirmed": true,
"credits_purchased": 1000,
"cost": "$300",
"current_balance": 1023,
"auto_refill_enabled": true,
"refill_threshold": 100,
"refill_amount": 500,
"next_billing_date": "2025-04-29"
}

Comparison Table

┌─────────────────────────────────────────────────────────────────┐
│ TIER COMPARISON │
├─────────────────────────────────────────────────────────────────┤
│ Feature │ Free │ Starter │ Builder │ Advanced │
├──────────────────────────┼────────┼─────────┼─────────┼──────────┤
│ Price │ $0 │ $29 │ $99 │ $499 │
│ Req/min │ 10 │ 100 │ 500 │ 5000 │
│ Operators │ 42 │ 234 │ 234 │ 234 │
│ Max error │ ≤1% │ ≤0.5% │ ≤0.1% │ ≤0.01% │
│ Advanced protocols │ 0% │ 70% │ 100% │ 100% │
│ ZeqProof verification │ Yes │ Yes │ Yes │ Yes │
│ Publication-ready │ No │ Yes │ Yes │ Yes │
│ Regulatory-compliant │ No │ No │ Yes │ Yes │
│ Team collaboration │ No │ No │ No │ Yes │
│ Dedicated support │ No │ No │ No │ Yes │
│ Custom SLA │ No │ No │ No │ Yes │
└─────────────────────────────────────────────────────────────────┘

Upgrading and Downgrading

Upgrading

curl -X POST https://zeq.dev/api/account/upgrade \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"new_tier": "builder"
}'

Response:

{
"upgrade_successful": true,
"previous_tier": "starter",
"new_tier": "builder",
"effective_date": "2025-03-29T13:30:00Z",
"prorated_refund": "$15.33",
"new_billing_date": "2025-04-29",
"next_invoice": {
"amount": "$83.67",
"description": "Builder tier (18 days remaining in cycle)"
}
}

Upgrades take immediate effect.

Downgrading

curl -X POST https://zeq.dev/api/account/downgrade \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"new_tier": "starter"
}'

Downgrades take effect at the next billing cycle to protect your workflows.

Cross-References

The Degradation Philosophy

The term "degradation" is intentional. Zeq's design reflects a principle:

Free-tier capabilities are a strict subset of paid-tier capabilities. Every upgrade strictly increases what you can do.

This is different from some platforms where free and paid tiers are "different" but not comparable. At Zeq:

Free ⊂ Starter ⊂ Builder ⊂ Advanced ⊂ Architect

(where ⊂ means "is a strict subset of")

Everything you can do on Free, you can do on Starter (plus more). Everything on Starter, you can do on Builder, etc.

Why This Philosophy?

  1. Fairness: You're not learning one thing on Free and discovering on Starter that your code doesn't transfer
  2. Scalability: As your project grows, you upgrade—your code doesn't break
  3. No Vendor Lock-in: You choose to upgrade because you need more capability, not because the platform forces you

Graceful Degradation: What Happens at Limits

Example 1: Free Tier, Hitting Rate Limit

You're prototyping with QM1 on Free tier. You write a loop that makes 15 computations per minute:

for i in range(15):
result = zeq.compute("QM1", {"energy_level": i})
print(result)

What happens:

  • Computation 1-10: ✓ Succeed
  • Computation 11: ✗ 429 Too Many Requests
  • Computation 12-15: ✗ All rejected with same error

Not degraded: The API doesn't slow down or reduce precision. It either works (✓) or doesn't (✗).

How to handle:

import time
for i in range(15):
try:
result = zeq.compute("QM1", {"energy_level": i})
except RateLimitError:
time.sleep(10) # Wait before retrying
result = zeq.compute("QM1", {"energy_level": i})

Or upgrade to Starter tier (100 req/min).

Example 2: Free Tier, Requesting Unavailable Operator

You try to use QM23 (advanced quantum mechanics):

curl -X POST https://zeq.dev/api/zeq/compute \
-H "Content-Type: application/json" \
-d '{"operators": ["QM23"]}'

Response:

{
"error": "OPERATOR_NOT_ACCESSIBLE",
"tier": "free",
"requested_operator": "QM23",
"reason": "Available only on Starter tier and above",
"status": 403
}

Not degraded: You don't get a lesser version of QM23. You get rejected cleanly with a helpful error.

How to handle:

  • Upgrade to Starter ($29/month), or
  • Use QM1 (available on Free), which covers basic quantum mechanics

Example 3: Starter Tier, Precision at Limit

You request a computation with very tight tolerance:

{
"operators": ["QM1"],
"precision": {
"relative_error_tolerance": 1e-8 // Very tight
}
}

Response (Starter tier):

{
"precision_validated": true,
"requested_tolerance": 1e-8,
"achieved_tolerance": 5e-4, // Only achieved this
"warning": "Starter tier precision limit is ≤0.5% (5e-4). Your request was tightened to this value.",
"recommendation": "Upgrade to Builder tier (≤0.1%) or Advanced (≤0.01%) for tighter tolerances."
}

Graceful degradation: The computation succeeds, but with the best precision your tier allows. You're not blindsided—the response tells you exactly what you got.

Example 4: Builder Tier Computation Hitting Precision Limit

You're doing regulatory work (FDA submission) on Builder tier and need ≤0.05% error:

{
"operators": ["QM1", "NM19"],
"precision": {
"relative_error_tolerance": 5e-4 // 0.05%
}
}

Response:

{
"result": {...},
"metadata": {
"achieved_error": 0.087, // ✓ Within tier limit
"tier_limit": 0.1,
"margin": 0.013, // Some headroom
"regulatory_qualified": true // Safe for FDA
}
}

Success: You achieved your precision goal within your tier. You can confidently submit to FDA.


Tier Upgrade Examples

Scenario 1: Student to Researcher

Starting point: Free tier

  • 10 computations/minute
  • 42 basic operators
  • ≤1% error
  • Use case: Learning quantum mechanics

After 6 months: Starter tier ($29/month)

  • 100 computations/minute (10× more)
  • 234 operators (5.5× more)
  • ≤0.5% error (2× tighter)
  • Use case: Publishing research paper

Code change: NONE. Your QM1 computation that worked on Free still works on Starter, just faster and more precise.

Scenario 2: Startup to Enterprise

Year 1: Builder tier ($99/month)

  • 500 computations/minute
  • All 234 operators
  • ≤0.1% error
  • Use case: Production drug simulation app

Year 3: Advanced tier ($499/month)

  • 5000 computations/minute (10× more)
  • Dedicated API endpoint (lower latency)
  • Team collaboration (multiple users)
  • Use case: Company-wide research platform

Year 5: Architect tier (custom)

  • Unlimited computations
  • ≤0.001% error
  • Custom on-premises deployment
  • Dedicated account manager
  • Use case: National lab with 100+ researchers

Code change: NONE (at the API level). The request format is identical. Only the performance and limits scale up.


When to Upgrade (Decision Matrix)

SituationAction
You hit rate limit more than once a weekUpgrade immediately
You need an operator not in your tierUpgrade (or wait for next billing cycle)
You're publishing results but precision warnings appearUpgrade (precision affects reputation)
You have a team (>1 person)Upgrade to Advanced (team features)
You're doing regulatory work (FDA, EMA)Upgrade to Builder minimum (≤0.1% error required)
You're prototyping and learningStay on Free (no payment, no risk)
You're publishing but don't hit precision limitsStarter is sufficient (but check edge cases)
You're running a small production serviceBuilder tier is standard choice
You're running a large research labAdvanced or Architect tier

See Also