LLMBenchmark.ai

Claimed by its makers

LLMBenchmark.ai is an LLM benchmark leaderboard that tracks 358 AI models across 206 public benchmarks, with API prices, updated daily.·llmbenchmark.ai

Browse all Analytics apps

Visit

Free See pricing ↓

LLMBenchmark.ai is a leaderboard and comparison site that aggregates public benchmark results for 358 large language models from 56 makers, ranking them overall and by category (coding, agentic, reasoning, math, knowledge, web design). It computes its own composite scores (the LLMBenchmark Index and Vibe Coding Index) from third-party benchmark data, pulls API pricing from OpenRouter, and derives a 'cost per solved task' metric. It offers side-by-side model comparisons, curated 'best for' lists, a pricing calculator, embeddable charts, a downloadable dataset, and a free API/MCP server for querying model data.

What it's for

Ranking LLMs overall and by category (coding, agentic, reasoning, math, knowledge, web design)Comparing any two models side by sideFinding the cheapest model that still performs well for a taskCalculating cost per solved coding taskComparing API prices across providersTracking newly released and trending modelsEmbedding leaderboard charts and comparison tables on other sitesQuerying model data programmatically via API or MCP server

Features 15

API pricing comparison by provider with a cost calculator
Best-for lists (cheapest model, best open-weight, best per maker, etc.)
Cost per solved task metric (SWE-bench Verified cost divided by resolve rate)
Daily price snapshots tracking price changes over time
Embeddable charts and comparison tables (CC BY 4.0 licensed)
Free API and MCP server for querying model data (no key, 500 requests/IP/day)
Full dataset download in JSON
LLMBenchmark Index: weighted overall score across 7 inputs
Methodology documentation detailing scoring weights and coverage rules
Model update changelog with RSS feed
New and trending models tracking (last 30 days, Hugging Face trending)
Search across 358 models, makers and comparisons
Separate Agentic, Reasoning, Math, Knowledge and Web design indexes
Side-by-side model comparison pages
Vibe Coding Index for coding ability

At a glance

free tierAPI

TypeWeb app

DeploymentCloud

PlatformsWeb

LanguagesEnglish

CompanyTake Group sp. z o.o. · 2019

Integrations

OpenRouterHugging Face

Pricing

The site itself is free to use; it does not sell subscriptions. All dollar prices shown on the pages (e.g. $0.05–$600 per million tokens) are third-party LLM API prices sourced from OpenRouter, not charges for LLMBenchmark.ai itself.