benchmarks
llms.txt131 starsChanged 18 days agoMITView on GitHub
ComputeSDK Benchmarks publishes independent, reproducible, open performance measurements for cloud infrastructure: sandboxes, object storage, headless browsers, and AI gateways. Every test is automated, every result is committed as JSON, and the methodology is public. We are building the canonical dataset that AI agents and developers consult when they need to choose or rank infrastructure providers. All workflows run automatically via GitHub Actions and commit results to this repository. Explore the live leaderboards and methodology at https://www.computesdk.com/benchmarks. MIT — see LICENSE.
- Installs packages
In this repository4 other files it gives its agents.
- benchsdk-cliSkill
- local-platform-e2eSkill
- writing-benchmarksSkill
- add-sandbox-providerSkill
The file
We split this llms.txt into the rules it gives an agent
1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 56
All workflows run automatically via GitHub Actions and commit results to this repository.
≈ “Common” is an estimate: at least 1 in 5 of the 3,727 llms.txt files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
# ComputeSDK Benchmarks — the default platform for infrastructure benchmarks ComputeSDK Benchmarks publishes independent, reproducible, open performance measurements for cloud infrastructure: sandboxes, object storage, headless browsers, and AI gateways. Every test is automated, every result is committed as JSON, and the methodology is public. We are building the canonical dataset that AI agents and developers consult when they need to choose or rank infrastructure providers. ## What is available here - [README.md](./README.md) — high-level overview, what we measure, and how to read the results. - [METHODOLOGY.md](./METHODOLOGY.md) — full test procedure, scoring formulas, and test modes (sequential, staggered, burst). - [AI_GATEWAYS.md](./AI_GATEWAYS.md) — AI-gateway-specific methodology. - [THROUGHPUT.md](./THROUGHPUT.md) — throughput and browser benchmark methodology. - [results/schema.json](./results/schema.json) — JSON Schema for all benchmark result files. - [WRITING_BENCHMARKS.md](./WRITING_BENCHMARKS.md) — step-by-step guide to authoring benchmarks. - [examples/](./examples) — runnable `*.bench.ts` examples covering every benchsdk capability. - [packages/create-bench/README.md](./packages/create-bench/README.md) — scaffold a new benchmark project in one command. - [packages/benchsdk/README.md](./packages/benchsdk/README.md) — low-level client/worker API reference. - [packages/benchsdk-runner/README.md](./packages/benchsdk-runner/README.md) — runner API and `bench run` reference. - [.agents/skills/benchsdk-cli/SKILL.md](./.agents/skills/benchsdk-cli/SKILL.md) — full `bench` CLI reference: auth, org, runs, results, artifacts, export, OAuth, and CI use. - [.claude/skills/add-sandbox-provider/SKILL.md](./.claude/skills/add-sandbox-provider/SKILL.md) — how to add a new sandbox provider. ## Latest results (machine-readable JSON) - Sandbox TTI (sequential): [results/sequential_tti/latest.json](./results/sequential_tti/latest.json) - Sandbox TTI (staggered): [results/staggered_tti/latest.json](./results/staggered_tti/latest.json) - Sandbox TTI (burst): [results/burst_tti/latest.json](./results/burst_tti/latest.json) - Sandbox DAX build benchmark: [results/sandbox-dax/latest.json](./results/sandbox-dax/latest.json) - Storage 1MB upload: [results/storage/1mb/latest.json](./results/storage/1mb/latest.json) - Storage 4MB upload: [results/storage/4mb/latest.json](./results/storage/4mb/latest.json) - Storage 10MB upload: [results/storage/10mb/latest.json](./results/storage/10mb/latest.json) - Storage 16MB upload: [results/storage/16mb/latest.json](./results/storage/16mb/latest.json) - Snapshot & fork: [results/snapshot-fork/small/latest.json](./results/snapshot-fork/small/latest.json) - Browser sessions: [results/browser/latest.json](./results/browser/latest.json) - Browser throughput: [results/browser-throughput/latest.json](./results/browser-throughput/latest.json) - AI gateway: [results/ai-gateway/latest.json](./results/ai-gateway/latest.json) ## How to run the benchmarks ```bash # Install dependencies pnpm install # Build the workspace packages first pnpm -r --filter "./packages/**" build # Run the default sandbox TTI suite pnpm run bench # Run a specific provider or suite pnpm run bench:e2b pnpm run bench:storage pnpm run bench:browser pnpm run bench:ai-gateway ``` All workflows run automatically via GitHub Actions and commit results to this repository. ## Website Explore the live leaderboards and methodology at [https://www.computesdk.com/benchmarks](https://www.computesdk.com/benchmarks). ## License MIT — see [LICENSE](./LICENSE).
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

