lm-evaluation-harness
by Community
A framework for few-shot evaluation of language models.
OSS
lm-evaluation-harness
Added 1 June 2026
Overview
Python framework for evaluating language models across standardized benchmarks using few-shot prompting. Supports multiple model backends and task definitions, enabling reproducible performance measurement against established datasets like MMLU, HellaSwag, and others.
Best for
Best for
Researchers and engineers benchmarking LLM performance against established academic standards
Use cases
- Comparing performance across different LLM architectures on standard benchmarks
- Measuring model degradation or improvement after fine-tuning or quantization
- Validating model behavior on specific task categories before deployment
Notes
Python framework for evaluating language models across standardized benchmarks using few-shot prompting. Supports multiple model backends and task definitions, enabling reproducible performance measurement against established datasets like MMLU, HellaSwag, and others.
12,772 stars on GitHub. Last updated 2026-05-11. Licensed MIT.
Use cases
- Comparing performance across different LLM architectures on standard benchmarks
- Measuring model degradation or improvement after fine-tuning or quantization
- Validating model behavior on specific task categories before deployment
Pros
- Extensive built-in benchmark library reduces setup time for common evaluations
- Supports multiple model backends (local, API-based, custom implementations)
- Active community maintenance with 12k+ stars and regular benchmark additions
Cons
- Steep learning curve for custom task definition and evaluation logic
- Evaluation runs can be computationally expensive and time-consuming at scale
- Limited guidance on interpreting results or statistical significance testing
Indexed from awesome-llm and enriched against its public facts.
Pros
- Extensive built-in benchmark library reduces setup time for common evaluations
- Supports multiple model backends (local, API-based, custom implementations)
- Active community maintenance with 12k+ stars and regular benchmark additions
Cons
- Steep learning curve for custom task definition and evaluation logic
- Evaluation runs can be computationally expensive and time-consuming at scale
- Limited guidance on interpreting results or statistical significance testing
Open-source & AI alternatives
Swap-in tools that solve the same job. Weigh the trade-offs before you commit.
OpenAI Evals
Community
Evals is a framework for evaluating LLMs and LLM systems, and an open-source registry of benchmarks.
promptfoo
Community
Test your prompts, agents, and RAGs. Red teaming/pentesting/vulnerability scanning for AI. Compare performance of GPT, Claude, Gemini, DeepSeek, and more. Simple declarative config
Ragas
Community
Supercharge Your LLM Application Evaluations 🚀
AlpacaEval
Community
AlpacaEval Leaderboard
Arize-Phoenix
Community
Arize Phoenix: Open Source AI Development Platform
Berkeley Function-Calling Leaderboard
Community
Explore The Berkeley Function Calling Leaderboard (also called The Berkeley Tool Calling Leaderboard) to see the LLM
Beyond the Imitation Game: Quantifying and extrapolating the capabilities of language models
Community
Beyond the Imitation Game collaborative benchmark for measuring and extrapolating the capabilities of language models
Chain-of-Thought Hub
Community
Benchmarking large language models' complex reasoning ability with chain-of-thought prompting
CompassRank
Community
评测榜单旨在为大语言模型和多模态模型提供全面、客观且中立的得分与排名，同时提供多能力维度的评分参考，以便用户能够更全面地了解大模型的能力水平。
HELM
Community
Holistic Evaluation of Language Models (HELM) is an open source Python framework created by the Center for Research on Foundation Models (CRFM) at Stanford for holistic, reproducib
InfiBench
Community
IInfiBench: Evaluating the Question-Answering Capabilities of Code LLMs
instruct-eval
Community
This repository contains code to quantitatively evaluate instruction-tuned models such as Alpaca and Flan-T5 on held-out tasks.
LangWatch
Community
The platform for LLM evaluations and AI agent testing
LawBench
Community
LawBench
lighteval
Community
Lighteval is your all-in-one toolkit for evaluating LLMs across multiple backends
LLMEval
Community
LLMEval is a research series dedicated to building comprehensive, fair, and robust evaluation frameworks for large language models.
M3CoT
Community
Leaderboard | M 3 CoT
MathEval
Community
a comprehensive benchmarking platform designed to evaluate large models' mathematical abilities across 20 fields and nearly 30,000 math problems.
MMedBench
Community
Medical Multilingual Benchmark
OLMO-eval
Community
Evaluation suite for LLMs
OpenAI Evals
Community
Evals is a framework for evaluating LLMs and LLM systems, and an open-source registry of benchmarks.
Ragas
Community
Supercharge Your LLM Application Evaluations 🚀
simple-evals
Community
Eval tools by OpenAI.
Pairs with
Other entries in the index that connect to this one. Click through to see the chain.
ACLUE
Community
Official github repo for ACLUE, an evaluation benchmark focused on ancient Chinese language comprehension
awesome-hallucination-detection
Community
List of papers on hallucination detection in LLMs.
awesome-language-model-analysis
Community
This paper list focuses on the theoretical and empirical analysis of language models, especially large language models (LLMs). The papers in this list investigate the learning beha
Awesome-LLM-hallucination
Community
LLM hallucination paper list
Awesome LLM Human Preference Datasets
Community
A curated list of Human Preference Datasets for LLM fine-tuning, RLHF, and eval.
Chinese Large Model Leaderboard
Community
非线智能 NoneLinear - ReLE评测：中文AI大模型能力评测（持续更新）：目前已囊括374个大模型，覆盖chatgpt、gpt-5.4、谷歌gemini-3.1-pro、Claude-4.6、文心ERNIE-X1.1、ERNIE-5.0、qwen3.6-max、qwen3.6-plus、百川、讯飞星火、商汤senseChat等商用模型， 以及st
CompMix
Community
CompMix: A Benchmark for Heterogeneous Question Answering.
EasyEdit
Community
[ACL 2024] An Easy-to-use Knowledge Editing Framework for LLMs.
FELM
Community
FELM: Benchmarking Factuality Evaluation of Large Language Models
Finetuned Language Models are Zero-Shot Learners
Community
This paper explores a simple method for improving the zero-shot learning abilities of language models. We show that instruction tuning—finetuning language models on a collection
GPT-NeoX
Community
An implementation of model parallel autoregressive transformers on GPUs, based on the Megatron and DeepSpeed libraries
Megatron-DeepSpeed
Community
Ongoing research training transformer language models at scale, including: BERT & GPT-2
Meta Lingua
Community
Meta Lingua: a lean, efficient, and easy-to-hack codebase to research LLMs.
MMToM-QA
Community
Leaderboard for the MMToM-QA benchmark (Jin et al., ACL 2024).
Neurips2022-Foundational Robustness of Foundation Models
Community
NeurIPS Tutorial Foundational Robustness of Foundation Models
OLMo-7B
Community
Artifacts for the first set of OLMo models.
PubMedQA
Community
PubMedQA Homepage
Pythia-1|1.4|2.8|6.9|12B
Community
The hub for EleutherAI's work on interpretability and learning dynamics
SuperLim
Community
a Swedish language understanding benchmark that evaluates natural language processing (NLP) models on various tasks such as argumentation analysis, semantic similarity, and textual
TAT-DQA
Community
TAT-DQA: A Document Visual Question Answering (VQA) Dataset, aiming to answer questions over visually-rich documents with a hybrid of Tabular and Textual Content in Finance
VisualWebArena
Community
Project webpage for the VisualWebArena paper.
WHOOPS!
Community
Breaking Common Sense: WHOOPS! A Vision-and-Language Benchmark of Synthetic and Compositional Images
Get the free Developer’s Field Guide
A 27-page field guide to the AI coding workflow with Claude. Claude Code, MCP servers, the prompt patterns that work, and what to delegate. Free.
Enter your work email. We send it straight over, plus a few short notes worth knowing. Unsubscribe any time.