Simple, Usage-Based Pricing
From $0.02 / job, pay only for what you use
No setup fees. Three tiers by model class, open-weight models only, billed per job in USD.
Estimate your bill
Drag your monthly volume. We pick the cheapest way to pay.
Every option and price below is one this page sells. The first $5 of usage on a new key is free.
Pay-per-job
Routed to a verified GPU on the network. Same OpenAI-compatible API for every tier.
Get started instantly. No credit card required.
- ✓ ~250 free small-model jobs ($5 credit)
- ✓ All open-source models
- ✓ OpenAI-compatible API
- ✓ Shared GPU pool
- ✓ Community support
Llama 3.2, Phi-4, DeepSeek-R1 7B, Qwen 2.5 7B. Fast, cheap, perfect for chat and routing.
- ✓ Models up to 8B params
- ✓ OpenAI-compatible API
- ✓ Pay-as-you-go
Qwen 2.5 14B (always-warm), 13–14B class. The best speed/quality tradeoff for production.
- ✓ Models 8–32B params
- ✓ Priority routing
- ✓ Always-warm vLLM workers
Qwen 3.6 27B, 32B+ reasoning models. Compete with Claude Sonnet at a fraction of the cost.
- ✓ Models 32B+ params
- ✓ Reasoning + long context
- ✓ Dedicated lane
Volume packs & subscriptions
Save up to 40% with prepaid packs. Or subscribe for monthly credits + priority routing.
10,000 small-model jobs. 40% off list. Credits never expire.
$25 of credit each cycle + priority routing. Cancel anytime.
- ✓ $25 credit / month (rolls over)
- ✓ Priority routing
- ✓ ADK / Python / Node SDKs
- ✓ Overage at list price
$200 credit + dedicated lane + 25% off overage. For production workloads.
- ✓ $200 credit / month
- ✓ Dedicated GPU lane
- ✓ 25% off overage
- ✓ Email support
Dedicated GPU capacity, KYC/KYT compliance, invoice billing.
- ✓ Dedicated GPU capacity
- ✓ Private model deployment
- ✓ KYC/KYT via Infinihash
- ✓ Dedicated support
Every pack and plan above can also be paid with MYAI — MYAI payment opens at TGE.
Pricing comparison
Cost per 1,000 medium-class jobs (8–32B). MyAi vs centralized providers.
| Provider | Cost per 1K jobs | Decentralized | No KYC required |
|---|---|---|---|
| MyAi Network — Medium† | $60 (1,000-job pack) · $80 at list | ✓ | ✓ |
| MyAi Network — Small† | $15 (1,000-job pack) · $20 at list | ✓ | ✓ |
| OpenAI gpt-4o-mini | ~$50 | ✗ | ✗ |
| Anthropic Claude Haiku | ~$80 | ✗ | ✗ |
| AWS Bedrock | $200–$2,000 | ✗ | ✗ |
† MyAi rows show the prepaid 1,000-job pack price sold at billing.infinihash.com (25% below list: Small $0.02 × 1,000 = $20; Medium $0.08 × 1,000 = $80). Competitor figures are approximate public list rates.
MyAi vs OpenAI: honest side-by-side
We win on privacy, decentralization, and open weights. OpenAI wins on frontier closed models and cold-start latency. No overclaiming.
| Dimension | MyAi Network | OpenAI |
|---|---|---|
| Privacy posture | Inference runs on the operator's hardware. No central log of prompts or outputs. | Data retained 30 days by default. API opt-out in org settings. |
| Training opt-out | N/A — we never see your data. | Default off via API; can be re-enabled via ChatGPT settings. |
| Data residency | Operator-dependent (global). Filter operators by region. Not guaranteed without vetting operators. | Azure datacenters (US / EU / Asia). Region-select on Enterprise. |
| Signed model releases | ✓ Ed25519 offline signing. Public key mirrored on-chain + GitHub + docs. | ✗ Closed weights. No public signing scheme. |
| Latency p50 | ~180 ms (Medium, warm node, US-East) | ~150 ms (gpt-4o-mini) |
| Latency p95 | ~3–8 s (cold GPU spin-up; varies by operator load) | <2 s |
| Pricing | $0.02 Small · $0.08 Medium · $0.25 Large per completion. Billed in USD; MYAI payment planned at TGE. | $0.15/$0.60 per M tokens (gpt-4o-mini) · $2.50/$10 per M tokens (gpt-4o) |
| On-prem / self-host | ✓ One-command node install. Earn MYAI per completed job. | ✗ Azure-only. |
| Open weights | ✓ All models are open-weight (Llama, Qwen, DeepSeek, Mistral). | ✗ GPT-4o, o1, o3 are closed. No downloadable weights. |
| Frontier closed models | ✗ Not available. We do not host GPT-5 or o-series. | ✓ GPT-4o, GPT-4.1, o1, o3. State-of-the-art on public evals. |
| API compatibility | ✓ Drop-in OpenAI-compatible. One env var: OPENAI_BASE_URL | ✓ Native |
† p50/p95 from US-East nodes, May 2026. Cold-start p95 varies with operator GPU load. Full comparison →
Ready to start?
Start with $5 of free credit, or run a node and earn MYAI per verified job from day one.