# Cloud Run GPUs vs Thunder Compute

> Cloud Run GPUs scores 68.4 (B) to Thunder Compute's 51.3 (D) for gpu compute. Prices, MCP, x402, uptime and agent notes side by side.

- Canonical: https://www.anchorterminal.com/compare/cloud-run-gpus-vs-thunder-compute
- Markdown: https://www.anchorterminal.com/compare/cloud-run-gpus-vs-thunder-compute.md (~5,150 tokens)
- Site index for agents: https://www.anchorterminal.com/llms.txt (the key docs and every listing in one file, at most 16 MB, about 4 million tokens: https://www.anchorterminal.com/llms-full.txt)
- API: https://www.anchorterminal.com/api/v1/index.json
- MCP: https://www.anchorterminal.com/mcp (the whole catalogue as tools, no key)
- DNS: dig +short TXT <slug>.grade.anchorterminal.com (the grade as a DNS TXT record, for a sandbox with no HTTP where names still resolve; also top., best.<category>. and <a>.vs.<b>. under the zone, and the same under grade.letme.dev)
- Licence: CC BY 4.0, credit Anchor Terminal and link this page (https://www.anchorterminal.com/terms/#data)
- Updated: 2026-11-07

Cloud Run GPUs scores 68.4 (B) on agent readiness against Thunder Compute's 51.3 (D), and leads in 6 of 7 scored categories. Both do gpu compute.

- Cloud Run GPUs: grade B, 68.4/100, rank #288 of 1486. Markdown https://www.anchorterminal.com/tools/cloud-run-gpus.md · JSON https://www.anchorterminal.com/api/v1/tools/cloud-run-gpus.json
- Thunder Compute: grade D, 51.3/100, rank #1024 of 1486. Markdown https://www.anchorterminal.com/tools/thunder-compute.md · JSON https://www.anchorterminal.com/api/v1/tools/thunder-compute.json
- Best GPU and serverless compute for AI workloads: https://www.anchorterminal.com/best/gpu-compute/index.md
- All 231 gpu compute comparisons: https://www.anchorterminal.com/compare/gpu-compute/index.md

## Which one, for what

### Cloud Run GPUs (B)

Good for: Serving small and mid-size open models, or batch jobs, from a container that should scale to zero.

Ahead on:
- Reliability, 70 against 65
- Schema & documentation, 81 against 70
- Agent ergonomics, 66 against 48
- Security & auth, 78 against 51
- Transparency & trust, 84 against 66

Also in its favour:
- No incidents deducted, where Thunder Compute loses 5 points for them

Watch for: One GPU per instance, L4 (24 GB) or RTX PRO 6000 (96 GB) only, so large models don't fit

### Thunder Compute (D)

Good for: Agents in coding tools that rent a single persistent GPU machine for development, fine-tuning or a model server, at low hourly prices and over MCP.

Watch for: No rate limits, 429 guidance or SLA were found in the reviewed documentation, and the terms disclaim availability


## Why the edge

### Security & auth: Cloud Run GPUs +27

- Cloud Run GPUs (78): OAuth 2.0 tokens from service accounts under IAM. Google's APIs accept `access_token` as a query parameter, the 10-point deduction other Google API listings take (20). IAM roles separate viewing, deploying and invoking, and services can require IAM authentication for every call (16). The platform runs your own container and returns its output (10). Cloud Audit Logs record Admin API calls, and request logs go to Cloud Logging (15). Google's vulnerability reward programme, security.txt and certifications (17). 20+16+10+15+17 = 78.
- Thunder Compute (51): The MCP server uses OAuth 2.0 with discovery, refresh tokens and ten scopes that separate read from write for instances, snapshots and sandboxes. The REST API uses named API tokens that are created and revoked one at a time, with no scopes found. No secret travels in a URL (25 of 30). Read scopes exist, and the CLI requires `--yes` to delete without a prompt. No confirmation step for `delete_instance` or `delete_snapshot` was found in the MCP docs, and both are described as irreversible (8 of 20). Responses are infrastructure metadata, except `run_command` output and sandbox logs, with no guidance found on untrusted output (8 of 15). `GET /tokens` returns creation and last-used times and sandbox logs are listed by API. No audit log was found (4 of 15). No security.txt, disclosure policy or bug bounty was found. `llms.txt` states SOC 2 Type II and GDPR compliance with no report or trust page linked, and the DPA promises an audit summary on written request and incident notice within 72 hours (6 of 20). The site readings are per the 9 October check, not re-read on 4 November. 25+8+8+4+6 = 51.

### Agent ergonomics: Cloud Run GPUs +18

- Cloud Run GPUs (66): Platform reading. One `gcloud run deploy` with `--gpu` and `--gpu-type` adds a GPU to a container, and responses carry the service object with `fieldMask` on updates (15). List calls page with `pageSize` and `pageToken` (15). Google's status codes and messages, with quota errors that name the limit (12). Revisions are immutable, traffic can be moved back, and etags guard updates, but no idempotency key on create (12). Client libraries in many languages and gcloud, though a GPU deploy needs instance-based billing, minimum CPU and memory and, past 3 GPUs, a quota request (12). 15+15+12+12+12 = 66.
- Thunder Compute (48): No field selection, and `GET /instances/list` returns every instance in one map. The MCP server has 28 tools in one set, with no toolsets found (12 of 25). `limit` up to 100, `page_token` and a status filter on sandboxes and sandbox logs. No pagination on instances, snapshots or keys (8 of 20). Errors carry a numeric `code`, a string `error` such as `invalid_request` and a `message`, documented per operation (14 of 20). Duplicate SSH keys are a no-op and port changes are deduplicated in one transaction, but create calls have no idempotency key. MCP tool annotations were not read (7 of 20). A create call needs five fields. The official Python SDK covers sandboxes only, and the Go CLI covers instances (7 of 15). All of this is per the 9 October check, not re-read on 4 November. 12+8+14+7+7 = 48.

### Transparency & trust: Cloud Run GPUs +18

- Cloud Run GPUs (84): Closed service under the Google Cloud terms, with Apache-2.0 client libraries (15). The Cloud Data Processing Addendum and the privacy notice cover customer data, per the Document AI listing's 1 November reading; no Cloud Run statement on GPU instance data after shutdown was found (22). The Cloud terms promise 12 months' notice before a backwards-incompatible API change, and driver versions are announced in the release notes (16). Regions per GPU type listed, with Google's subprocessor list and data-location terms (16). 15+22+16+16 = 69.
- Thunder Compute (66): Closed service under Terms and Conditions dated 28 September 2026 and Delaware law. The CLI is MIT. The OpenAPI document names Apache 2.0 and a placeholder terms address (15 of 30). The privacy policy of 28 September 2026 says instance data is not used to train AI models, the terms say instance data may be deleted 60 days after the account was last active, and a DPA is published. The sub-processor list names Firebase for sign-in while the MCP docs name Stytch, and the FAQ says instances are EC2 while the list names Azure (18 of 30). No deprecation policy. The terms allow changes without notice, and the billing docs say rates can change without notice. Legacy endpoints are marked as retained for compatibility (6 of 20). Five sub-processors with locations are listed and the DPA sets 10 business days' notice of changes. The list is dated 27 August 2025 and omits Stytch and Sentry, which the MCP docs and the CLI source name (12 of 20). The site readings are per the 9 October check, not re-read on 4 November. 15+18+6+12 = 51.

### Schema & documentation: Cloud Run GPUs +11

- Cloud Run GPUs (81): API reading. A public Discovery document for the Cloud Run Admin API v2 (25). No llms.txt; Google's docs answer Markdown only at an unlinked `.md.txt` address, read as half, as for the Document AI listing (5). The GPU page says when to turn zonal redundancy on or off and the best-practices page when to bake models into images or load them from Cloud Storage (16). The GPU type is a free string in a node selector, with CPU and memory minimums stated in prose (9). gcloud, YAML, Terraform and console steps on each page, and errors follow Google's status model with no Cloud Run GPU error list (11). Versioned API (v1 and v2) and dated release notes (15). 25+5+16+9+11+15 = 81.
- Thunder Compute (70): Public OpenAPI 3.1 document at api.thundercompute.com:8443/openapi.json with 41 operations and 72 schemas (25). `llms.txt`, `llms-full.txt` and a Markdown copy of every docs page (10). 28 of 41 operations have a description. Instance operations state side effects such as the private key being returned once, while most sandbox operations have a summary only (12 of 20). Required fields are marked, but the document has six enums, `gpu_type` and `template` are free strings and the instance list is a map keyed by ID (8 of 15). 12 examples and one error schema with `code`, `error` and `message`. Most operations document 400, 401 and 500, and none documents 429 (9 of 15). `/v1` and `/v2` paths, with legacy endpoints and the `mode` field marked as retained or deprecated. No API changelog was found. The CLI has a dated one (6 of 15). No published copy of the MCP tool definitions was found, so they were not read. All of this is per the 9 October check, not re-read on 4 November. 25+10+12+8+9+6 = 70.

### Reliability: Cloud Run GPUs +5

- Cloud Run GPUs (70): Hosted reading. Google Cloud's status page with a JSON incident feed (20). In the 90 days from 8 August 2026 the feed names Cloud Run in three multi-product incidents, a service outage on 20 August (15:40 to 19:20 UTC, 27 products, global and us-west1), a network degradation in us-central1-b on 1 September and a europe-west2 disruption on 20 October. One major (10). Resource limits and GPU quota are published with numbers (3 GPUs a region by default), but no request rate for the Admin API was found on the quotas page (10). Google's client libraries retry with backoff and the Admin API takes etags, while no GPU-specific retry guidance was found (10). An SLA for GPU services, 99.95% with zonal redundancy and 99.5% without (10). GPU services generally available since 7 April 2025 (10). 20+10+10+10+10+10 = 70.
- Thunder Compute (65): Graded on the hosted lines for the REST API and the hosted MCP server. Thunder Compute's terms bar robots, scrapers and data-gathering tools on the Services, so its site, status page and API host weren't requested on 4 November and the readings below are per the 9 October check, not re-read on 4 November. Status page on incident.io at status.thundercompute.com with Website, API and Instances components and monthly history (20). No incidents listed for July to 9 October 2026. The rest of the window is unread and not credited, so the line keeps its recorded award (30). No rate limits with numbers were found in the docs or the OpenAPI document (0). No 429 or retry guidance in the docs. The CLI source reads a `Retry-After` header on errors, adding a duplicate SSH key is a no-op and port updates are atomic, but instance creation has no idempotency key (5 of 15, a partial value kept and flagged). No SLA found, and the terms say availability is not guaranteed (0). The API and MCP server carry no beta label (10). 20+30+0+5+0+10 = 65.

## Score by category

| Category | Weight | Cloud Run GPUs | Thunder Compute | Edge |
| --- | --- | --- | --- | --- |
| Reliability | 16% (20 this run) | 70 | 65 | Cloud Run GPUs +5 |
| Performance | 10%, pending | pending | pending | not scored in this run |
| Schema & documentation | 13% (16.2 this run) | 81 | 70 | Cloud Run GPUs +11 |
| Agent ergonomics | 13% (16.2 this run) | 66 | 48 | Cloud Run GPUs +18 |
| Security & auth | 14% (17.5 this run) | 78 | 51 | Cloud Run GPUs +27 |
| Payments & pricing | 10% (12.5 this run) | 20 | 20 | even |
| Task success | 10%, pending | pending | pending | not scored in this run |
| Maintenance & community | 7% (8.8 this run) | 80 | 79 | Cloud Run GPUs +1 |
| Transparency & trust | 7% (8.8 this run) | 84 | 66 | Cloud Run GPUs +18 |
| Negative events | ≤15 | 0 | -5 | |
| **Total** | | **68.4 · B** | **51.3 · D** | |

## Facts side by side

| Fact | Cloud Run GPUs | Thunder Compute |
| --- | --- | --- |
| Kind | Platform | HTTP API |
| Vendor | Google Cloud | Thunder Compute |
| Hosted endpoint | `https://run.googleapis.com` | `https://api.thundercompute.com:8443/v1` |
| Transports |  | HTTP, Streamable HTTP |
| Auth | OAuth | OAuth or key |
| Pricing | Pay per use | Pay per use |
| Price for gpu compute | not published | $0.219 per GB per month |
| x402 | no | no |
| Licence | Proprietary service under the Google Cloud Platform terms. Client libraries are Apache-2.0 | Proprietary service under Thunder Compute's Terms and Conditions. The `tnr` CLI on GitHub is MIT |
| Tools exposed | none | 28 |
| Read-only variant documented | no | no |
| llms.txt | no | yes |
| MCP registry | not listed | `io.github.Thunder-Compute/thunder-compute` |
| Last release | 2026-10-30 | 2026-10-27 |
| Terms last updated | 2026-09-02 | 2026-09-28 |
| Privacy policy last updated | 2026-09-28 | 2026-09-28 |
| Customer content may train models | not found in the text | not found in the text |
| Terms restrict automated access | not found in the text | not found in the text |
| Terms restrict benchmarking | not found in the text | yes |
| Terms or service can change without notice | not found in the text | yes |
| Arbitration or class-action waiver | not found in the text | yes |
| Popularity | none | 34 stars |

## How the scores have moved

### Cloud Run GPUs

No score change since it was listed.

### Thunder Compute

- -5.0 on 4 November 2026 (C to D): negative event
- +0.2 on 27 October 2026: checklist

## Verdicts

**Cloud Run GPUs.** Serverless GPUs on Cloud Run, an L4 or RTX PRO 6000 per instance, billed per second, scaling to zero and starting in about 5 seconds before the model loads, with a published SLA and IAM. One GPU per instance, 3 GPUs of default quota a region and instance-based billing limit it to small and mid-size models.

**Thunder Compute.** The hosted MCP server signs in with OAuth and separate read and write scopes, and the REST API has a public OpenAPI 3.1 document. No rate limits, SLA or API changelog were found, instance creation has no idempotency key, and the CLI sends error reports to Sentry by default with no documented opt-out.

## Before you call either

### Cloud Run GPUs

1. Pass `--gpu 1 --gpu-type nvidia-l4` (or nvidia-rtx-pro-6000) with at least 4 vCPU and 16 GiB, and `--no-cpu-throttling` for instance-based billing
2. Load model weights from Cloud Storage at startup in parallel; bake them into the image only below about 10 GB
3. Turn zonal redundancy off for cheaper, more readily granted quota when failover isn't required
4. Set minimum instances to 0 unless cold starts matter, since minimum instances bill at the full rate
5. Send OAuth tokens in the `Authorization` header, never as `access_token` in the URL

### Thunder Compute

1. Call `GET /v2/status` or the `get_availability` tool before creating an instance. Availability can change before launch, and creation fails when a type is sold out.
2. List instances before retrying a failed create, because the call has no idempotency key.
3. Pass `public_key` on create. If omitted, the response carries a generated private key that is returned once.
4. To pause work, create a snapshot, delete the instance and later create a new instance from the snapshot. Snapshot storage keeps billing until deleted.
5. For headless use set `TNR_API_TOKEN` to a token from the console, and upgrade the CLI to v2.1.0 or later, which stops copying that credential onto instances.

## Questions

### Which is better for AI agents, Cloud Run GPUs or Thunder Compute?

Cloud Run GPUs scores 68.4 (B) on agent readiness against Thunder Compute's 51.3 (D), and leads in 6 of 7 scored categories.

### Can an agent call Cloud Run GPUs and Thunder Compute without installing anything?

Yes. Cloud Run GPUs has a hosted endpoint at https://run.googleapis.com and Thunder Compute at https://api.thundercompute.com:8443/v1.


## For agents

- This comparison as JSON: https://www.anchorterminal.com/compare/cloud-run-gpus-vs-thunder-compute.json, and with the fewest tokens: https://www.anchorterminal.com/compare/cloud-run-gpus-vs-thunder-compute.min.md
- Over MCP at https://www.anchorterminal.com/mcp (no key): `compare_tools {"a": "cloud-run-gpus", "b": "thunder-compute"}`. From a terminal: `anchor compare cloud-run-gpus thunder-compute`
- Each listing in full: https://www.anchorterminal.com/api/v1/tools/cloud-run-gpus.json and https://www.anchorterminal.com/api/v1/tools/thunder-compute.json

## Other comparisons with Cloud Run GPUs or Thunder Compute

- [Baseten vs Cloud Run GPUs](https://www.anchorterminal.com/compare/baseten-vs-cloud-run-gpus.md)
- [Baseten vs Thunder Compute](https://www.anchorterminal.com/compare/baseten-vs-thunder-compute.md)
- [Beam vs Cloud Run GPUs](https://www.anchorterminal.com/compare/beam-vs-cloud-run-gpus.md)
- [Beam vs Thunder Compute](https://www.anchorterminal.com/compare/beam-vs-thunder-compute.md)
- [Cerebrium vs Cloud Run GPUs](https://www.anchorterminal.com/compare/cerebrium-vs-cloud-run-gpus.md)
- [Cerebrium vs Thunder Compute](https://www.anchorterminal.com/compare/cerebrium-vs-thunder-compute.md)
- [Cloud Run GPUs vs CoreWeave](https://www.anchorterminal.com/compare/cloud-run-gpus-vs-coreweave.md)
- [Cloud Run GPUs vs Crusoe Cloud](https://www.anchorterminal.com/compare/cloud-run-gpus-vs-crusoe-cloud.md)
- [Cloud Run GPUs vs GPUsMarket](https://www.anchorterminal.com/compare/cloud-run-gpus-vs-gpusmarket.md)
- [Cloud Run GPUs vs Hugging Face Inference Endpoints](https://www.anchorterminal.com/compare/cloud-run-gpus-vs-hugging-face-inference-endpoints.md)
- [Cloud Run GPUs vs Hyperbolic](https://www.anchorterminal.com/compare/cloud-run-gpus-vs-hyperbolic.md)
- [Cloud Run GPUs vs Lambda Cloud](https://www.anchorterminal.com/compare/cloud-run-gpus-vs-lambda.md)
- [Cloud Run GPUs vs Massed Compute](https://www.anchorterminal.com/compare/cloud-run-gpus-vs-massed-compute.md)
- [Cloud Run GPUs vs Nebius AI Cloud](https://www.anchorterminal.com/compare/cloud-run-gpus-vs-nebius-ai-cloud.md)
- [Cloud Run GPUs vs Northflank](https://www.anchorterminal.com/compare/cloud-run-gpus-vs-northflank.md)
- [Cloud Run GPUs vs Together GPU Clusters](https://www.anchorterminal.com/compare/cloud-run-gpus-vs-together-gpu-clusters.md)
- [CoreWeave vs Thunder Compute](https://www.anchorterminal.com/compare/coreweave-vs-thunder-compute.md)
- [Crusoe Cloud vs Thunder Compute](https://www.anchorterminal.com/compare/crusoe-cloud-vs-thunder-compute.md)
- [GPUsMarket vs Thunder Compute](https://www.anchorterminal.com/compare/gpusmarket-vs-thunder-compute.md)
- [Hugging Face Inference Endpoints vs Thunder Compute](https://www.anchorterminal.com/compare/hugging-face-inference-endpoints-vs-thunder-compute.md)
- [Hyperbolic vs Thunder Compute](https://www.anchorterminal.com/compare/hyperbolic-vs-thunder-compute.md)
- [Koyeb vs Thunder Compute](https://www.anchorterminal.com/compare/koyeb-vs-thunder-compute.md)
- [Lambda Cloud vs Thunder Compute](https://www.anchorterminal.com/compare/lambda-vs-thunder-compute.md)
- [Massed Compute vs Thunder Compute](https://www.anchorterminal.com/compare/massed-compute-vs-thunder-compute.md)
- [Modal vs Thunder Compute](https://www.anchorterminal.com/compare/modal-vs-thunder-compute.md)
- [Nebius AI Cloud vs Thunder Compute](https://www.anchorterminal.com/compare/nebius-ai-cloud-vs-thunder-compute.md)
- [Northflank vs Thunder Compute](https://www.anchorterminal.com/compare/northflank-vs-thunder-compute.md)
- [Replicate Deployments vs Thunder Compute](https://www.anchorterminal.com/compare/replicate-deploy-vs-thunder-compute.md)
- [Runpod vs Thunder Compute](https://www.anchorterminal.com/compare/runpod-vs-thunder-compute.md)
- [Thunder Compute vs Together GPU Clusters](https://www.anchorterminal.com/compare/thunder-compute-vs-together-gpu-clusters.md)
- [Thunder Compute vs Vast.ai](https://www.anchorterminal.com/compare/thunder-compute-vs-vast-ai.md)
- [Thunder Compute vs Verda](https://www.anchorterminal.com/compare/thunder-compute-vs-verda.md)
- [Cloud Run GPUs vs Koyeb](https://www.anchorterminal.com/compare/cloud-run-gpus-vs-koyeb.md)
- [Cloud Run GPUs vs Modal](https://www.anchorterminal.com/compare/cloud-run-gpus-vs-modal.md)
- [Cloud Run GPUs vs Replicate Deployments](https://www.anchorterminal.com/compare/cloud-run-gpus-vs-replicate-deploy.md)
- [Cloud Run GPUs vs Runpod](https://www.anchorterminal.com/compare/cloud-run-gpus-vs-runpod.md)
- [Cloud Run GPUs vs Vast.ai](https://www.anchorterminal.com/compare/cloud-run-gpus-vs-vast-ai.md)
- [Cloud Run GPUs vs Verda](https://www.anchorterminal.com/compare/cloud-run-gpus-vs-verda.md)

---

Query this catalogue directly over MCP: https://www.anchorterminal.com/mcp (Streamable HTTP, no key). How: https://www.anchorterminal.com/agents/#mcp