Free; MIT license. You pay your LLM provider for the tokens a benchmark consumes.
First released
2026
Last verified
Answer
I am speed is a fast.com for LLM APIs: open the page, paste your own API key, hit Run, and watch tokens stream in with live tokens per second, time to first token, and time to last token. Everything runs in the browser against the provider directly — nothing is proxied through a server and.
About
I am speed is a fast.com for LLM APIs: open the page, paste your own API key, hit Run, and watch tokens stream in with live tokens per second, time to first token, and time to last token. Everything runs in the browser against the provider directly — nothing is proxied through a server and keys are stored locally under AES-GCM encryption — so the numbers reflect the path your own client would take. OpenAI, Anthropic, Groq, Cerebras, Fireworks AI, Mistral, OpenRouter, Google Gemini, x.ai, z.ai, and local servers such as Ollama, LM Studio, and llama.cpp are supported, with model discovery from the endpoint for local setups (which must allow browser CORS, and need a tunnel or a local build to avoid mixed-content blocking on the hosted site). Race Mode runs the same prompt against two or three providers in parallel and animates them on a racetrack, deciding the winner by TTLT with tokens per second as tiebreaker and a separate award for the lowest TTFT. Results can be submitted to a public leaderboard backed by DynamoDB for provider and model comparisons. The Astro app is MIT licensed and runs locally with npm for private benchmarking.
Best for
Token-level latency and throughput benchmarks against LLM inference endpoints
Teams that prefer open-source licensing and self-hosting options
Groups comfortable scripting in None
Not the best fit when
Workloads outside the listed protocols or languages without a proof-of-concept
Architecture and concurrency
I am speed is commonly operated as hybrid software with scenario authoring in None. Concurrency and distribution depend on how you size workers or injectors.
Specs
Scripting
None
Protocols
HTTPSSSE
Operating systems
Browser
Tags
aillmbrowserweb-uiquick-start
CATALOG LAST VERIFIED · 2026-08-12
Frequently asked
ANSWER-SHAPED
Who is I am speed for?
I am speed is best for teams benchmarking LLM and generative AI endpoints on token latency and throughput. It is open source, so teams can self-host and extend it. Vendor: NaveenKumar Namachivayam. Deployment model: hybrid. Free; MIT license. You pay your LLM provider for the tokens a benchmark consumes.
How does I am speed differ from AIPerf?
Compared with AIPerf, deployment is hybrid versus self-hosted for AIPerf; scripting centers on None rather than Python, YAML. I am speed lists protocols HTTPS, SSE; AIPerf lists HTTP, HTTPS, SSE. Status is active versus active. Use the Test Rig to compare full specs side by side before a proof-of-concept.
What scripting language and concurrency model does I am speed use?
I am speed uses None for scripting or scenario definition. Runtime model: protocol-oriented virtual users configured through the product workflow. OS support: Browser. Protocols: HTTPS, SSE. Match these to how your team already authors tests and which systems you must drive under load.
What is the cloud or enterprise path for I am speed?
I am speed supports a hybrid path that can combine self-hosted controllers or agents with cloud or distributed load as the product allows. Teams often keep sensitive assets private while bursting generators into the cloud. As open source, commercial support or hosted siblings may exist separately. Check the official site. Pricing note: Free; MIT license. You pay your LLM provider for the tokens a benchmark consumes.
What are known limitations of I am speed?
I am speed limitations to weigh: fit depends on team skills, protocol needs, CI integration, and total cost of ownership rather than feature checklists alone. Status: active. Always validate against your target protocols and scale with a proof-of-concept. Confirm details against the official site and a short proof-of-concept in your environment.