OpenCompass ORG

OpenCompass is an LLM evaluation platform, supporting a wide range of models (Llama3, Mistral, InternLM2,GPT-4,LLaMa2, Qwen,GLM, Claude, etc) over 100+ datasets.

Open-source evaluation toolkit of large multi-modality models (LMMs), support 220+ LMMs, 80+ benchmarks

Official Repo of "MMBench: Is Your Multi-modal Model an All-around Player?"

[EMNLP 2025] CompassVerifier: A Unified and Robust Verifier for LLMs Evaluation and Outcome Reward

The All-in-one Judge Models introduced by Opencompass

Official repo of "MMBench-GUI: Hierarchical Multi-Platform Evaluation Framework for GUI Agents". It can be used to evaluate a GUI agent with a hierarchical manner across multiple platforms, including Windows, Linux, macOS, iOS, Android and Web.

A toolkit for inference and evaluation of 'mixtral-8x7b-32kseqlen' from Mistral AI

Benchmarking Legal Knowledge of Large Language Models

[ACL2024] T-Eval: Evaluating Tool Utilization Capability of Large Language Models Step by Step

Evaluating LLMs' multi-round chatting capability via assessing conversations generated by two LLM instances.

[NeurIPS 2024 D&B] GTA: A Benchmark for General Tool Agents & [arXiv 2026] GTA-2

A Comprehensive Benchmark for Software Development.

No description.

[ACL 2024 Findings] MathBench: A Comprehensive Multi-Level Difficulty Mathematics Evaluation Dataset

No description.

[ACL 2024] ANAH & [NeurIPS 2024] ANAH-v2 & [ICLR 2025] Mask-DPO

The official implementation of "Ada-LEval: Evaluating long-context LLMs with length-adaptable benchmarks"

[NeurIPS 2024] A comprehensive benchmark for evaluating critique ability of LLMs

The first unified, efficient, and extensible evaluation toolkit for evaluating image generation and editing models across multiple benchmarks.

[ACL 2025] Are Your LLMs Capable of Stable Reasoning?

[EMNLP 2024 Findings] ProSA: Assessing and Understanding the Prompt Sensitivity of LLMs

A multi-language code evaluation tool.

Assessing Context-Aware Creative Intelligence in MLLMs

We provide TextEdit, a high-quality, multi-scenario text editing benchmark for generation models.

CNFinBench — the first comprehensive benchmark for high-stakes financial scenarios. It spans 29 subtasks grounded in authoritative financial corpora and real business contexts, reconstructing end-to-end agent execution chains from requirement parsing, path planning, tool invocation, to result verification.

Official Repo of "CIBench: Evaluation of LLMs as Code Interpreter "

[ICLR 2026] Rectifying LLM Thought From Lens of Optimization

The code repository for the NeurIPS 2025 paper "Rethinking Verification for LLM Code Generation: From Generation to Testing."

Demo data of CompassBench

No description.

[Preprint 2025] Deciphering Trajectory-Aided LLM Reasoning: An Optimization Perspective

ATLAS: A High-Difficulty, Multidisciplinary Benchmark for Frontier Scientific Reasoning

No description.

No description.

No description.

Code for the paper "Evaluating Large Language Models Trained on Code"

No description.

No description.

Sphinx Theme for OpenCompass - Modified from PyTorch

EvalPlus for rigourous evaluation of LLM-synthesized code

No description.

No description.

No description.

No description.

MiroMind Research Agent: Fully Open-Source Deep Research Agent with Reproducible State-of-the-Art Performance on FutureX, GAIA, HLE, BrowserComp and xBench.

Deep Research Agent CognitiveKernel-Pro from Tencent AI Lab. Paper: https://arxiv.org/pdf/2508.00414

No description.

No description.

A clean documentation and blog theme for your Hugo site based on Bootstrap 5

No description.