# AI Models & Benchmarks

Where to compare models: leaderboards, the benchmarks behind them, and each model family's official page.

Source page: https://ai.justout.today/models/  
Last walkthrough: Nov 20, 2026, 12:03 AM PT

## Leaderboards

### LMArena
Crowd-voted head-to-head model rankings.  
Link: https://arena.ai  
Checked: Nov 20, 2026, 12:03 AM PT

### Artificial Analysis
Independent comparisons of model quality, speed and price.  
Link: https://artificialanalysis.ai  
Checked: Nov 20, 2026, 12:03 AM PT

### LiveBench
Benchmark with fresh questions to limit contamination.  
Link: https://livebench.ai  
Checked: Nov 20, 2026, 12:03 AM PT

### Aider LLM leaderboards
How well models edit code in Aider's polyglot benchmark.  
Link: https://aider.chat/docs/leaderboards/  
Checked: Nov 20, 2026, 12:03 AM PT

### HELM
Stanford's holistic evaluation of language models.  
Link: https://crfm.stanford.edu/helm/  
Checked: Nov 20, 2026, 12:03 AM PT

### MTEB leaderboard
Massive Text Embedding Benchmark rankings.  
Link: https://huggingface.co/spaces/mteb/leaderboard  
Checked: Nov 20, 2026, 12:03 AM PT

## Reasoning & knowledge benchmarks

### MMLU
57-subject multiple-choice knowledge test.  
Link: https://arxiv.org/abs/2009.03300  
Checked: Nov 20, 2026, 12:03 AM PT

### GPQA
Graduate-level, 'Google-proof' science questions.  
Link: https://arxiv.org/abs/2311.12022  
Checked: Nov 20, 2026, 12:03 AM PT

### Humanity's Last Exam
Very hard expert-written questions across many fields.  
Link: https://lastexam.ai  
Checked: Nov 20, 2026, 12:03 AM PT

### ARC-AGI
Abstract reasoning puzzles easy for people, hard for AI.  
Link: https://arcprize.org  
Checked: Nov 20, 2026, 12:03 AM PT

### FrontierMath
Unpublished research-level maths problems.  
Link: https://epoch.ai/frontiermath  
Checked: Nov 20, 2026, 12:03 AM PT

### MATH
Competition mathematics problems with step-by-step solutions.  
Link: https://arxiv.org/abs/2103.03874  
Checked: Nov 20, 2026, 12:03 AM PT

### GSM8K
Grade-school maths word problems.  
Link: https://github.com/openai/grade-school-math  
Checked: Nov 20, 2026, 12:03 AM PT

### MMMU
College-level multimodal understanding and reasoning.  
Link: https://mmmu-benchmark.github.io  
Checked: Nov 20, 2026, 12:03 AM PT

### SimpleQA
Short factual questions to measure hallucination.  
Link: https://openai.com/index/introducing-simpleqa/  
Checked: bot-blocked

## Coding & agent benchmarks

### SWE-bench
Can a model resolve real GitHub issues?  
Link: https://www.swebench.com  
Checked: Nov 20, 2026, 12:03 AM PT

### HumanEval
Classic Python function-writing benchmark.  
Link: https://github.com/openai/human-eval  
Checked: Nov 20, 2026, 12:03 AM PT

### Terminal-Bench
Agents completing real tasks in a terminal.  
Link: https://www.tbench.ai  
Checked: Nov 20, 2026, 12:03 AM PT

### tau-bench
Tool-using agents in customer-service style tasks.  
Link: https://github.com/sierra-research/tau-bench  
Checked: Nov 20, 2026, 12:03 AM PT

### OSWorld
Agents operating real desktop computers.  
Link: https://arxiv.org/abs/2404.07972  
Checked: Nov 20, 2026, 12:03 AM PT

### WebArena
Agents completing tasks on realistic websites.  
Link: https://webarena.dev  
Checked: Nov 20, 2026, 12:03 AM PT

### GAIA
General AI assistant tasks needing tools and browsing.  
Link: https://arxiv.org/abs/2311.12983  
Checked: Nov 20, 2026, 12:03 AM PT

### BrowseComp
Hard-to-find facts that require persistent web browsing.  
Link: https://openai.com/index/browsecomp/  
Checked: bot-blocked

## Model families (official pages)

### OpenAI models
Current OpenAI model list and capabilities.  
Link: https://developers.openai.com/api/docs/models  
Checked: Nov 20, 2026, 12:03 AM PT

### Claude models
Current Claude model list.  
Link: https://platform.claude.com/docs/en/models/overview  
Checked: Nov 20, 2026, 12:03 AM PT

### Gemini models
Current Gemini API models.  
Link: https://aistudio.google.com/docs/models  
Checked: Nov 20, 2026, 12:03 AM PT

### Gemma
Google's open-weight model family.  
Link: https://deepmind.google/models/gemma/  
Checked: Nov 20, 2026, 12:03 AM PT

### Llama
Meta's open-weight model family.  
Link: https://dev.meta.ai/  
Checked: Nov 20, 2026, 12:03 AM PT

### Mistral models
Mistral's open and commercial models.  
Link: https://docs.mistral.ai/getting-started/models/  
Checked: Nov 20, 2026, 12:03 AM PT

### Qwen
Alibaba's open model family on GitHub.  
Link: https://github.com/QwenLM  
Checked: Nov 20, 2026, 12:03 AM PT

### DeepSeek
DeepSeek's open model releases on GitHub.  
Link: https://github.com/deepseek-ai  
Checked: Nov 20, 2026, 12:03 AM PT

## Research trackers

### Epoch AI
Data on AI compute, models and trends.  
Link: https://epoch.ai  
Checked: Nov 20, 2026, 12:03 AM PT

### METR
Measures how long a task AI agents can finish on their own.  
Link: https://metr.org  
Checked: Nov 20, 2026, 12:03 AM PT

