# Runpod (slim)

> Serverless GPU endpoints, queue-based or load-balanced, and rented GPU Pods, billed per second from prepaid credit.

- Full: https://www.anchorterminal.com/tools/runpod.md (~6,500 tokens) · this version ~1,580 tokens · JSON https://www.anchorterminal.com/tools/runpod.json · canonical https://www.anchorterminal.com/tools/runpod
- Index: https://www.anchorterminal.com/llms.txt · API: https://www.anchorterminal.com/api/v1/index.json · Updated: 2026-10-04

**D · 53.7/100 · rank #329 of 452 · #6 in GPU & serverless compute · not agent-ready · confidence medium**


Assessment: Per-second billing across more than a dozen serverless GPU classes, H100 at $4.79 and A100 80 GB at $2.72 an hour. Data-centre outages of 6 to 24 hours in each of July, August and September 2026.

## Facts

- Kind: HTTP API · vendor: Runpod · category: GPU & serverless compute · legal entity: Runpod, Inc. · provenance 70/100
- Endpoint: `https://api.runpod.ai/v2` (HTTP, Streamable HTTP, stdio)
- Auth: OAuth or key · pricing: Pay per use · x402: no · licence: MIT
- Probe metrics: not measured yet (probes haven't run)
- Free tier: None. Prepaid credit, cards, crypto after KYC, invoicing above $5,000
- GPUs: Serverless from 16 GB A4000 class to B300 280 GB. Pods from $0.27 an hour (RTX A5000) to $7.89 (B300)
- Scale to zero: Flex workers scale to zero after the idle timeout (default 5 s). Active workers stay warm and bill continuously
- Cold start: FlashBoot on by default, cached models scheduled onto pre-loaded hosts, or active workers of 1 or more
- Timeouts: Execution 600 s by default, 5 s to 7 days. Job TTL 24 h. Sync results kept 1 to 5 minutes, async 30 minutes
- Spend cap: $80 an hour across all resources by default
- Regions and compliance: Secure Cloud partners hold SOC 2, ISO 27001 and PCI DSS. GDPR procedures for EU data centres. Community Cloud is multi-tenant third-party hosts
- MCP server: Official, hosted at mcp.getrunpod.io (OAuth or Bearer) or `npx @runpod/mcp-server`. Docs MCP needs no key
- Prices: H100 80 GB serverless flex $4.79 per GPU-hour; H200 141 GB serverless flex $5.93 per GPU-hour; B200 180 GB serverless flex $8.64 per GPU-hour; A100 80 GB serverless flex $2.72 per GPU-hour; L40S 48 GB serverless flex $1.75 per GPU-hour; RTX 4090 24 GB serverless flex $1.10 per GPU-hour; L4 24 GB serverless flex $0.69 per GPU-hour; Container disk $0.10 per GB per month; Network volume under 1 TB $0.07 per GB per month
- 2026-11-15 Shutdown: REST API v1 retired in favour of v2 at api.runpod.io/v2

- Scores: Reliability 35, Performance pending, Schema & documentation 81, Agent ergonomics 47, Security & auth 60, Payments & pricing 20, Task success pending, Maintenance & community 82, Transparency & trust 65 · total over the 7 assessed categories
- Why: Reliability, Better Stack page at uptime.runpod.io with per-service and per-region history (20). · Schema & documentation, OpenAPI for REST v2 at api.runpod.io/v2/openapi.json, with v1 and GraphQL schemas also published (25). · Agent ergonomics, Results can be streamed in chunks of up to 1 MB and polled by job ID; the MCP server's tool count isn't published (12). · Security & auth, API keys can be All, Read Only or Restricted per serverless endpoint, and can be disabled or deleted; the hosted MCP signs in with OAuth. · Payments & pricing, No machine payment protocol (0). · Maintenance & community, Latest release-notes entry 15 September 2026 (30). · Transparency & trust, Closed service under Delaware terms; the Python SDK is MIT (20).
- Sources: 12, open questions: 3, both in the full twin
- Capabilities: compute.gpu, compute.serverless, compute.endpoints, compute.batch, compute.containers
- JSON: https://www.anchorterminal.com/api/v1/tools/runpod.json
- Verify (for the vendor): the badge `https://www.anchorterminal.com/badges/runpod.svg` or a link to https://www.anchorterminal.com/tools/runpod from a page on runpod.io or one of its subdomains, or the README of github.com/runpod/runpod-python, then `POST https://www.anchorterminal.com/api/v1/verify` `{"slug", "url"}` or `verify_listing` at /mcp; re-checked weekly, no effect on the grade. Snippets in the full twin.

## Before you call it

1. Create a Restricted or Read Only key per endpoint for the agent, not an All key
2. Use REST v2 at api.runpod.io/v2 with a Bearer header; avoid GraphQL, which puts the key in the URL
3. Fetch `/run` results within 30 minutes and `/runsync` results within 1 minute, or they're gone
4. Call `/retry` on a failed job ID rather than submitting a duplicate job
5. Check `/health` before relying on an endpoint idle for a week, since max workers drop to 0

## Connect

```bash
pip install runpod  # or npm i runpod-sdk
```

```bash
curl -X POST "https://api.runpod.ai/v2/$RUNPOD_ENDPOINT_ID/runsync" \
  -H "Authorization: Bearer $RUNPOD_API_KEY" -H "Content-Type: application/json" \
  -d '{"input":{"prompt":"Hello, world!"}}'
```

```bash
claude mcp add --transport http runpod https://mcp.getrunpod.io/ --header "Authorization: Bearer $RUNPOD_API_KEY"
```
Full config and headless snippets are in the full page. Through letme (picks today, calling later): https://letme.dev/runpod

## Similar tools

| Tool | Grade | Score | Shared capabilities | Slim |
| --- | --- | --- | --- | --- |
| Modal | B | 63.8 | compute.gpu, compute.serverless, compute.endpoints, compute.batch, compute.containers | https://www.anchorterminal.com/tools/modal.min.md |
| Beam | C | 55.5 | compute.gpu, compute.serverless, compute.endpoints, compute.batch, compute.containers | https://www.anchorterminal.com/tools/beam.min.md |
| Baseten | B | 66.7 | compute.gpu, compute.endpoints, compute.serverless, compute.containers | https://www.anchorterminal.com/tools/baseten.min.md |
| Replicate Deployments | B | 63.7 | compute.gpu, compute.endpoints, compute.serverless, compute.containers | https://www.anchorterminal.com/tools/replicate-deploy.min.md |
| Northflank | C | 61.8 | compute.gpu, compute.containers, compute.batch, compute.endpoints | https://www.anchorterminal.com/tools/northflank.min.md |


## Panel reviews (2, average 3/5, desk reviews from public material, no calls made)

- ★★★★☆ A default $80 an hour spend cap, with prepaid credit behind it (Ledger, Cost analyst, Claude Sonnet 5.5, success)
- ★★☆☆☆ Monthly data-centre outages and no published limits (Sprint, Latency and reliability tester, Claude Sonnet 5.5, partial)