Start free on pay-as-you-go, or commit to your usage and unlock custom discounts.
Pay as you go with provider prices. No commitments.
$0 per month
Pay as you go
Key features:
Lower your rates with discounts. Based on a usage commitment.
Custom discount
and more models with discount.
All the features of Self-Serve, and:
Every request routes at the provider’s officially published rate. New models are added to LLM API on their launch day.
| Popular models | Input ($/1M) | Output ($/1M) | Context | EUR USD | |
|---|---|---|---|---|---|
DeepSeek V4.1 Flash |
Vision Tools Reasoning Streaming Web Search | $0.3/req | $1.2/req | 1M | Get started |
GPT-6 Astra |
Vision Tools Reasoning Streaming Web Search | $10/req | $50/req | 1.1M | Get started |
Gemini 3.8 Flash |
Vision Tools Reasoning Streaming Web Search | $0.75/req | $3.75/req | 1M | Get started |
Claude Fable 5.1 |
Vision Tools Reasoning Streaming Web Search | $10/req | $50/req | 1M | Get started |
Qwen3.8 Flash Next |
Vision Tools Reasoning Streaming Web Search | $0.201/req | $0.5/req | 262.1K | Get started |
GLM-5.3-Flash |
Vision Tools Reasoning Streaming Web Search | $0.15/req | $0.5/req | 1M | Get started |
GLM-5.3 |
Tools Reasoning Streaming Web Search | $1.4/req | $4.4/req | 1M | Get started |
Gemini 3.7 Flash |
Vision Tools Reasoning Streaming Web Search | $0.75/req | $3.75/req | 1M | Get started |
Grok 4.6 |
Vision Tools Reasoning Streaming Web Search | $2/req | $6/req | 500K | Get started |
Qwen 3.8 Max |
Vision Tools Reasoning Streaming Web Search | $2/req | $6/req | 1M | Get started |
DeepSeek V4 Flash 0731 |
Tools Reasoning Streaming Web Search | $0.14/req | $0.28/req | 1M | Get started |
Claude Opus 5 |
Vision Tools Reasoning Streaming Web Search | $5/req | $25/req | 1M | Get started |
Region options, private routing, migration from a self-hosted gateway.
Retention controls, zero-retention routes, audit logs.
Per-team attribution, budget limits, model policies, one invoice.
MSA and DPA, SLA, dedicated capacity, personalized support.
Only pay for the compute you use, down to the minute. Deploy using Custom Models.
| GPU/Instance | vRAM | vCPU | RAM | Hourly price | |
|---|---|---|---|---|---|
| Nvidia H200 | 141 GB | 23 cores | 256 GB | $5.00 | Deploy |
| 2x Nvidia H200 | 282 GB | 46 cores | 512 GB | $10.00 | Deploy |
| 4x Nvidia H200 | 564 GB | 92 cores | 1024 GB | $20.00 | Deploy |
| 8x Nvidia H200 | 1128 GB | 184 cores | 2048 GB | $40.00 | Deploy |
| Nvidia RTX PRO 6000 Blackwell | 96 GB | 23 cores | 256 GB | $2.75 | Deploy |
| 2x Nvidia RTX PRO 6000 Blackwell | 192 GB | 46 cores | 512 GB | $5.50 | Deploy |
| 4x Nvidia RTX PRO 6000 Blackwell | 384 GB | 92 cores | 1024 GB | $11.00 | Deploy |
| 8x Nvidia RTX PRO 6000 Blackwell | 768 GB | 188 cores | 2048 GB | $22.00 | Deploy |
| Nvidia A100 | 80 GB | 11 cores | 145 GB | $2.50 | Deploy |
| 2x Nvidia A100 | 160 GB | 22 cores | 290 GB | $5.00 | Deploy |
| 4x Nvidia A100 | 320 GB | 47 cores | 580 GB | $10.00 | Deploy |
| 8x Nvidia A100 | 640 GB | 95 cores | 1160 GB | $20.00 | Deploy |
| Nvidia L4 | 24 GB | 7 cores | 30 GB | $0.80 | Deploy |
| 4x Nvidia L4 | 96 GB | 48 cores | 185 GB | $3.80 | Deploy |
| Nvidia L40S | 48 GB | 7 cores | 30 GB | $1.80 | Deploy |
| 4x Nvidia L40S | 192 GB | 47 cores | 380 GB | $8.30 | Deploy |
| 8x Nvidia L40S | 384 GB | 190 cores | 1532 GB | $23.50 | Deploy |
| Nvidia H100 | 80 GB | 25 cores | 240 GB | $10.00 | Deploy |
| 2x Nvidia H100 | 160 GB | 51 cores | 480 GB | $20.00 | Deploy |
| 4x Nvidia H100 | 320 GB | 102 cores | 960 GB | $40.00 | Deploy |
| Nvidia T4 | 16 GB | 3 cores | 15 GB | $0.50 | Deploy |
| 4x Nvidia T4 | 64 GB | 46 cores | 192 GB | $3.00 | Deploy |
| Nvidia A10G | 24 GB | 6 cores | 30 GB | $1.00 | Deploy |
| 4x Nvidia A10G | 96 GB | 46 cores | 186 GB | $5.00 | Deploy |
Our platform meets the world’s leading security and compliance standards — SOC 2, ISO 27001, GDPR, and CCPA — so your data stays secure, private, and always audit-ready.
FAQ
LLM API uses pay-as-you-go pricing with no monthly fee or long-term commitment. You pay only for the model usage you consume at each provider’s officially published rate.
No. Requests are billed at the provider’s published price—never a cent more. Eligible provider promotions are also passed through.
You get access to more than 400 models from over 30 providers through one API. Newly released models are added on launch day whenever supported.
The self-serve plan includes unlimited API keys with custom limits, cost and team management, usage analytics, and access to all supported models.
Yes. Enterprise customers can receive custom discounts based on a usage commitment, a custom contract and SLA, dedicated onboarding, priority support, invoice billing, and custom routing or model deployments.
Yes. LLM API supports controls such as OIDC, zero-retention routes, retention settings, and audit logs. The platform states that it meets SOC 2, ISO 27001, GDPR, and CCPA standards.
