speedbench
speedbench is Paired A/B speed benchmark for OpenAI-compatible LLM servers: decode tok/s with confidence intervals, TTFT, concurrency, long context, multi-turn agent sessions - knapcio/speedbench. It is ranked #1499 on the Dev Radar, in Testing & observability, first seen 11 days ago and shared in 3 posts (3.8k views).
GitHub - knapcio/speedbench: Paired A/B speed benchmark for OpenAI-compatible LLM servers: decode tok/s with confidence intervals, TTFT, concurrency, long context, multi-turn agent sessions · GitHub Skip to content Navigation Menu Sign in Appearance settings Search / Sign in Sign up Appearance settings You signed in with another tab or window. Reload to refresh your session. You signed out in another tab or window. Reload to refresh your session. You switched accounts on another tab or window. Reload to refresh your session. Dismiss alert {{ message }} Repository navigation speedbench Public Notifications You must be signed in to change…
What people said about speedbench on X
Full GLM-5.3 (753B) TP4 on 4× DGX Spark on @mmastrac's gr8 engine: recipe v0.9 🐉 • Prose decode: 40.6 tok/s at c1 with thinking on (RigMark) • +40 % prose over upstream gr8 on the same boxes (29.0 → 40.6) • Code 60.6 / structured 75.5 tok/s at c1 • Aggregate: 47.0 / 59.9 / 76.6 tok/s at c1 / c2 / c4 • Prefill:…
— @majewskizby, 11 days ago · 47 likes · see the post
Releasing speedbench: a speed benchmark for local LLM servers, closer to how you actually use a model 📏 • 73 prompts: prose, code, structured, agentic tool calls, multilingual, long documents • thinking on, like real use • time to first token, multi-turn agent sessions, concurrency • paired A/B runs with a…
— @majewskizby, 11 days ago · 18 likes · see the post
🔥🔥🔥
— @mmastrac, 11 days ago · 10 likes · see the post
Alternatives to speedbench
- Introducing Vercel CI — Vercel CI is in Beta. It runs only the checks your change affects, on compute sized to the push, using the Turborepo…
- Dagger — After five years of v0, Dagger 1.0 is in beta: official modules, test-native CI, remote execution and native GitHub…
- e2e — 300 stars left for 10k
- x18n — x18n reads the locale files of any project, finds where each key is used, and reports missing, unused and broken keys.…
- jam.dev — Google Tag Manager, Google Analytics, Amplitude, AppsFlyer, Segment, Plausible, Tinybird, Meta, TikTok, X, Reddit,…
- Signal65 PINNACLE — Gemini 4 Argon leads Signal65 PINNACLE ( ) at its default effort, and at Google's standard $4/$20 rate a correct task…
speedbench in numbers
- Rank on the Dev Radar: #1499 of 8801
- Shared in 3 posts by 2 accounts: @majewskizby, @mmastrac
- 3.8k views on those posts
- First seen 11 days ago, last shared 11 days ago
- Pricing seen by AGK Intelligence: open source
- Market: Testing & observability
FAQ
What is speedbench?
Paired A/B speed benchmark for OpenAI-compatible LLM servers: decode tok/s with confidence intervals, TTFT, concurrency, long context, multi-turn agent sessions - knapcio/speedbench It was first shared on X 11 days ago and is ranked #1499 on the Dev Radar.
Is speedbench free?
It is open source.
Who shared speedbench?
2 accounts on X, including @majewskizby, @mmastrac, in 3 posts totalling 3.8k views.
Every post is read and classified by AGK Intelligence: what it is, which market it belongs to, and whether the link is a real tool. 31.2k posts from 5k X accounts over the last 21 days, 8.8k tools, 12 markets. New posts are collected several times an hour and the full ranking is recomputed every hour — last update 2026-10-30 21:12 UTC. Full method.