Skip to content
RunInfra logo

RunInfra

Serve open models to agents with one API for OpenAI and Anthropic SDKs

Paid plans only
Tracked since 2026

The Bottom Line

Price
Paid plans only. Compare plans

Key facts

  • Runs open models optimized for agent workloads, with session-aware routing that directs repeat calls to the warm replica to reach high cache-hit rates.
  • Single API key works with OpenAI and Anthropic SDKs, offering both per-token pay-as-you-go and subscription coding plans for agent clients.
  • Cached input tokens are billed at a lower rate, and the platform reports cache-hit rates per model so effective input cost is often far below list price.

Pros

  • High cache-hit rates reduce effective input costs for repeated agent context
  • Single API key works across OpenAI and Anthropic SDKs, simplifying integration
  • Offers both pay-as-you-go credits and monthly coding plans to fit different usage patterns

Cons

  • Cache hits are best effort rather than guaranteed
  • Self-serve pay-as-you-go tier has no contractual SLA

What is RunInfra?

RunInfra is a hosted inference platform that serves open model APIs specifically designed for AI agent workloads. It exposes one API key compatible with both the OpenAI and Anthropic SDKs, letting developers use chat completions and messages-style endpoints without changing their existing agent code. The platform includes a library of open models, each with published token pricing for input, cached input, and output. It is backed by Y Combinator and targets use cases where an agent repeatedly sends the same context, such as coding assistants and autonomous agent loops. The platform also offers GPU-style alternatives with a GPU cost comparison tool, plus subscription-based coding plans for CLI agents like Claude Code and opencode, with usage caps that reset on a rolling basis.

Key Features

  • OpenAI-compatible chat completions and Anthropic-compatible Messages endpoints through one key
  • Session-aware routing that keeps repeat calls on the warm replica holding the session context
  • Cached input tokens billed at a reduced per-token price on supported models
  • Coding plans for Claude Code, opencode, Codex, and over 20 other agents, with usage limits that reset every 5 hours and weekly
  • Free Standby access when a coding plan limit is reached
  • Per-token pricing with as little as a $5 credit top-up that stays valid for one year
  • Optional auto-recharge for prepaid balances
  • Streaming support and JSON mode on supported chat models

Pricing Plans

Pricing checked Nov 12, 2026. Source: runinfra.ai/pricing, read Nov 12, 2026

RunInfra plans and prices, checked November 2026
Plan Price Details
Pay as you go

$5 minimum top-up

  • Pay per token at published prices
  • Credits remain valid for one year
  • One prepaid balance
+1 more
  • No seats or per-user fees
Enterprise

Custom pricing

  • Dedicated infrastructure
  • Custom volume and contract terms
  • Audit logs and role-based access control
+3 more
  • SOC 2 Type II
  • Custom SLAs up to 99.99%
  • Dedicated customer success manager and private Slack

Is RunInfra worth the price?

Good value

Best for development teams already committed to AI agent workflows.

Hidden Costs & Gotchas

$5 minimum top-up on Pay as you go

Usage limits reset every 5 hours and weekly on Starter, Pro, Team

Full RunInfra pricing: every plan, hidden costs and alternatives

Email me when RunInfra changes its price

We email you when RunInfra changes its pricing. We read its pricing page every week. One email per change, unsubscribe any time.

Have an account? Sign in and it is on in one click.

Recent pricing updates

Changes we detected on the RunInfra pricing page.

  • Plans changed: Pay as you go went from no price to $5 a month; Starter, Pro, Team removed.

Reviews

Improve Your Thinking Patterns Using ChatGPT cover
$99 Free with your review

Write the first RunInfra review, get a free AI guide

Share your experience and we will send you Improve Your Thinking Patterns Using ChatGPT, free.

Write the first review

Best RunInfra Alternatives

Top alternatives based on features, pricing, and user needs.

Most buyers shortlist 2 or 3 tools before committing. Pull a side-by-side comparison or browse the full alternatives shortlist below.

Explore More

RunInfra FAQ

How does session-aware routing keep agent context warm?

RunInfra routes repeat calls from the same agent session to the replica that already holds that session's context. The first call is cold, but subsequent calls hit the warm cache on the same replica, which can lead to very high cache-hit rates across a multi-turn agent conversation.

Can I use RunInfra with the OpenAI and Anthropic SDKs at the same time?

Yes. RunInfra exposes both OpenAI-compatible chat completions at POST /v1/chat/completions and Anthropic-compatible Messages at POST /v1/messages, so one API key works with either SDK depending on the model and endpoint you use.

What is the difference between the pay-as-you-go balance and a coding plan?

Pay-as-you-go is a prepaid credit balance starting at $5 where you pay per token at published prices. A coding plan is a monthly subscription from $10 that includes a set amount of usage, with limits that reset every 5 hours and weekly; when the limit is reached, you can use credits or free Standby.

How does RunInfra report cached input tokens on usage bills?

Cache hits appear in usage and are billed at the cached input rate instead of the full input rate, on models that support cached input discounts. The platform publishes the measured cache-hit rate per model from the last 24 hours of billed agent traffic.

Which models does RunInfra host?

RunInfra hosts open models including Ornith 1.5 122B, DeepSeek V4.1 Flash 1017, GLM 5.3, GLM 5.3 Flash, Nemotron 3.5 Lightning 30B, Qwen3.8 27B, Ornith 1.5 35B, and DeepSeek V4.1 Flash. Each model page lists limits, modality capabilities, and per-token pricing.

Is there a free way to test the Model APIs before paying?

Yes, RunInfra gives $1 free credit to test the Model APIs. After that, you add credit top-ups starting at $5.

Does the coding plan stop working when its usage limit is reached?

No, not necessarily. When a coding plan hits its limit, you can continue using credits from your balance, or use free Standby which provides a limited amount of chat per day at reduced concurrency when the platform is busy.

What does RunInfra offer for enterprise deployments?

Enterprise customers get dedicated infrastructure, custom volume and contract terms, custom SLAs up to 99.99%, SOC 2 Type II attestation, audit logs, role-based access control, a dedicated customer success manager, and a private Slack channel.

Source: runinfra.ai

Guides & Articles