Skip to main content
Applies to v3.6.1Reviewed 2026-04-20

Pricing

Pricing is fixed monthly and scales by team size, compute, and isolation level — no per-token surprises.

Live retail prices and billing terms are on achiral.ai/pricing. This page is the reference for what each tier means, what's always included, and when to upgrade.

The 12 tiers

TierTeam sizeTarget
Spark1–3Solo founders, individual developers
Seed3–10Early-stage startups
Rise10–25Expanding startups
Boost25–50Established small businesses
Scale50–100Mid-market
Grow100–250Large companies, single department
Expand250–500Large companies, multiple departments
Prime500–1,000Enterprise
Public1,000–2,500Large enterprise, multiple divisions
Elite2,500–5,000Global corporations
Titan5,000–10,000Fortune 500, massive organizations
God10,000+Hyperscale enterprises

What every tier includes

  • Dedicated Weaviate tenant with logical isolation.
  • OpenAI-compatible REST and WebSocket APIs.
  • Chiro (organization assistant) plus personal Executive Assistants.
  • Memory, retrieval-augmented generation, the 12-connector catalog.
  • Agentic actions with policy and approval guardrails.
  • Audit logs, RBAC, SSO (Growth tiers and above).

How tiers differ

  • Quotas scale with tier: tokens per request (2k → 16k), tokens per hour, concurrent requests.
  • Fine-tuning. LoRA fine-tuning is Seed and above.
  • Advanced capabilities. Resource provisioning, external API access, and organization onboarding automation unlock on Grow and above.
  • Audit retention. 30 days (Spark) to 365+ days (Growth+, configurable).
  • Pod isolation. Elite and above migrate to dedicated vLLM pods in their own Kubernetes namespace.
  • BAAs. HIPAA-aligned BAAs are signed on Elite and above.
  • BYOK. Customer-managed encryption keys are Dedicated tier only.

When to upgrade

  • Your team passes the tier's size range.
  • You hit rate limits during normal work hours.
  • You need fine-tuning on your own data.
  • You need HIPAA-aligned BAAs or audit-log retention beyond 90 days.
  • You need pod isolation for compliance, performance, or contractual reasons.

Pod isolation triggers

Achiral auto-promotes an organization to dedicated vLLM pods when any of:

  • Tier is Elite, Titan, or God.
  • Monthly spend exceeds $10,000.
  • Sustained inference utilization ≥ 80% for seven consecutive days.
  • The customer's compliance or performance posture requires dedicated compute.

FAQ

Is there a free trial? Yes. Fourteen days at Spark-tier quotas, no credit card required.

Do you charge per token? No. Fixed monthly pricing. Quotas are rate limits, not cost meters.

Can I move down a tier? Yes, at the end of a billing cycle. Data, fine-tunes, and connectors persist.

Can I run on my own hardware? Colocation on customer-owned DGX Spark hardware is available Elite and above. Contact sales.

Are prices the same globally? Base price is USD. Regional pricing and currencies are handled at checkout.

Next steps