Llamabench

Llamabench
Crowd-sourced local LLM benchmarks for any hardware, showing exact configurations for optimal token/sec performance.
About
What does Llamabench do?
Llamabench is a crowd-sourced platform for benchmarking local Large Language Models (LLMs). It allows users to discover the exact hardware configurations and settings, such as quantization and KV cache, that achieve optimal tokens-per-second performance. The service serves GPU buyers, quant tinkerers, and inference engineers looking to understand and improve their local LLM rig performance.
What industry does Llamabench operate in?
Llamabench operates in Foundation Model, Large Language Model, AI Infrastructure, AI Testing, Developer Tools, Benchmarking.
Ashr provides a test and evaluation platform for AI agents, generating synthetic user stories to identify errors and ensure quality across various modalities.
An AI startup focused on improving agent readiness and discoverability for AI agents to interact with websites.
Seeing AI from the right angle. Free, open tools, resources, reports, and data that show what AI systems actually do.
A platform that enhances the accuracy of AI models by incorporating human input to mitigate errors and ensure precision in tasks such as drafting documents, content creation, and data aggregation.
Testados e avaliados nas melhores IAs do planeta. Economize tempo, dinheiro e energia.
Growgence is the AI visibility platform that measures how ChatGPT, Gemini, Perplexity and Claude mention, cite and describe your brand.
Nxlogic PRIME evaluates AI. NsY turns evaluated intelligence into executive action.
Thenorthmark provides orientation for AI, helping users navigate and understand AI systems.
An entrant is a company tagged AI Testing whose domain was first registered in the window, counted from registry records in the StartupHub directory. 4 of them registered in the last 30 days. Registry detection runs two to three weeks behind registration, so recent weeks are a floor.
No comments yet. Be the first to share your take.