Hugging Face's logo

Datasets:
alpha-one-index
/
awesome-ai-index

Tasks:
Text Classification
Question Answering
Languages:
English
Size:
1K<n<10K
ArXiv:
Tags:
ai
machine-learning
deep-learning
awesome-list
ai-tools
llm
License:
Dataset card Files
xet
 Community
Dataset Viewer
Duplicate
The dataset viewer is not available for this split.
Cannot load the dataset split (in streaming mode) to extract the first rows.
Error code:   StreamingRowsError
Exception:    CastError
Message:      Couldn't cast
$schema: string
metadata: struct<version: string, updated: timestamp[s], license: string>
  child 0, version: string
  child 1, updated: timestamp[s]
  child 2, license: string
benchmarks: list<item: struct<id: string, name: string, full_name: string, category: string, tasks: int64, max_s (... 50 chars omitted)
  child 0, item: struct<id: string, name: string, full_name: string, category: string, tasks: int64, max_score: int64 (... 38 chars omitted)
      child 0, id: string
      child 1, name: string
      child 2, full_name: string
      child 3, category: string
      child 4, tasks: int64
      child 5, max_score: int64
      child 6, higher_is_better: bool
      child 7, url: string
name: string
version: string
description: string
generated_at: string
sections: struct<models: list<item: struct<model_id: string, task: string, downloads: int64, likes: int64, cre (... 1544 chars omitted)
  child 0, models: list<item: struct<model_id: string, task: string, downloads: int64, likes: int64, created_at: string (... 623 chars omitted)
      child 0, item: struct<model_id: string, task: string, downloads: int64, likes: int64, created_at: string, last_modi (... 611 chars omitted)
          child 0, model_id: string
          child 1, task: string
          child 2, downloads: int64
          child 3, likes: int64
          child 4, created_at: string
          child 5, last_modified: string
          child 6, tags: list<item: string>
              child 0, item: string
          c
...
ild 1, name: string
                  child 2, full_name: string
                  child 3, category: string
                  child 4, tasks: int64
                  child 5, max_score: int64
                  child 6, higher_is_better: bool
                  child 7, url: string
  child 3, frameworks: list<item: struct<$schema: string, metadata: struct<version: string, updated: timestamp[s], license: (... 182 chars omitted)
      child 0, item: struct<$schema: string, metadata: struct<version: string, updated: timestamp[s], license: string>, f (... 170 chars omitted)
          child 0, $schema: string
          child 1, metadata: struct<version: string, updated: timestamp[s], license: string>
              child 0, version: string
              child 1, updated: timestamp[s]
              child 2, license: string
          child 2, frameworks: list<item: struct<id: string, name: string, jurisdiction: string, status: string, effective_date: ti (... 58 chars omitted)
              child 0, item: struct<id: string, name: string, jurisdiction: string, status: string, effective_date: timestamp[s], (... 46 chars omitted)
                  child 0, id: string
                  child 1, name: string
                  child 2, jurisdiction: string
                  child 3, status: string
                  child 4, effective_date: timestamp[s]
                  child 5, url: string
                  child 6, risk_levels: list<item: string>
                      child 0, item: string
to
{'name': Value('string'), 'description': Value('string'), 'generated_at': Value('string'), 'version': Value('string'), 'sections': {'models': List({'model_id': Value('string'), 'task': Value('string'), 'downloads': Value('int64'), 'likes': Value('int64'), 'created_at': Value('string'), 'last_modified': Value('string'), 'tags': List(Value('string')), 'pipeline_tag': Value('string'), 'fetched_at': Value('string'), '$schema': Value('string'), 'metadata': {'version': Value('string'), 'updated': Value('timestamp[s]'), 'source': Value('string'), 'license': Value('string')}, 'models': List({'id': Value('string'), 'name': Value('string'), 'vendor': Value('string'), 'parameters_b': Value('float64'), 'context_window': Value('int64'), 'release_date': Value('timestamp[s]'), 'benchmarks': {'mmlu': Value('float64'), 'gpqa_diamond': Value('float64'), 'humaneval': Value('float64'), 'swe_bench_verified': Value('float64'), 'math_500': Value('float64'), 'chatbot_arena': Value('int64')}, 'license_spdx': Value('string'), 'eu_ai_act_risk': Value('string'), 'ntia_compliant': Value('bool'), 'known_cves': List(Value('null')), 'model_type': Value('string'), 'modality': Value('string')})}), 'vendors': List({'$schema': Value('string'), 'vendors': List({'id': Value('string'), 'name': Value('string'), 'hq_country': Value('string'), 'founded': Value('int64'), 'funding_usd': Value('int64'), 'employees_est': Value('int64'), 'models_indexed': Value('int64'), 'primary_modality': Value('string'), 'eu_ai_act_risk': Value('string'), 'soc2': Value('bool'), 'iso27001': Value('bool'), 'gdpr': Value('bool'), 'ccpa': Value('bool'), 'website': Value('string')})}), 'benchmarks': List({'$schema': Value('string'), 'metadata': {'version': Value('string'), 'updated': Value('timestamp[s]'), 'license': Value('string')}, 'benchmarks': List({'id': Value('string'), 'name': Value('string'), 'full_name': Value('string'), 'category': Value('string'), 'tasks': Value('int64'), 'max_score': Value('int64'), 'higher_is_better': Value('bool'), 'url': Value('string')})}), 'frameworks': List({'$schema': Value('string'), 'metadata': {'version': Value('string'), 'updated': Value('timestamp[s]'), 'license': Value('string')}, 'frameworks': List({'id': Value('string'), 'name': Value('string'), 'jurisdiction': Value('string'), 'status': Value('string'), 'effective_date': Value('timestamp[s]'), 'url': Value('string'), 'risk_levels': List(Value('string'))})})}}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/src/services/worker/src/worker/utils.py", line 99, in get_rows_or_raise
                  return get_rows(
                         ^^^^^^^^^
                File "/src/libs/libcommon/src/libcommon/utils.py", line 272, in decorator
                  return func(*args, **kwargs)
                         ^^^^^^^^^^^^^^^^^^^^^
                File "/src/services/worker/src/worker/utils.py", line 77, in get_rows
                  rows_plus_one = list(itertools.islice(ds, rows_max_number + 1))
                                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 2690, in __iter__
                  for key, example in ex_iterable:
                                      ^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 2227, in __iter__
                  for key, pa_table in self._iter_arrow():
                                       ^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 2251, in _iter_arrow
                  for key, pa_table in self.ex_iterable._iter_arrow():
                                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 494, in _iter_arrow
                  for key, pa_table in iterator:
                                       ^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 384, in _iter_arrow
                  for key, pa_table in self.generate_tables_fn(**gen_kwags):
                                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/packaged_modules/json/json.py", line 265, in _generate_tables
                  self._cast_table(pa_table, json_field_paths=json_field_paths),
                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/packaged_modules/json/json.py", line 120, in _cast_table
                  pa_table = table_cast(pa_table, self.info.features.arrow_schema)
                             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2272, in table_cast
                  return cast_table_to_schema(table, schema)
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2218, in cast_table_to_schema
                  raise CastError(
              datasets.table.CastError: Couldn't cast
              $schema: string
              metadata: struct<version: string, updated: timestamp[s], license: string>
                child 0, version: string
                child 1, updated: timestamp[s]
                child 2, license: string
              benchmarks: list<item: struct<id: string, name: string, full_name: string, category: string, tasks: int64, max_s (... 50 chars omitted)
                child 0, item: struct<id: string, name: string, full_name: string, category: string, tasks: int64, max_score: int64 (... 38 chars omitted)
                    child 0, id: string
                    child 1, name: string
                    child 2, full_name: string
                    child 3, category: string
                    child 4, tasks: int64
                    child 5, max_score: int64
                    child 6, higher_is_better: bool
                    child 7, url: string
              name: string
              version: string
              description: string
              generated_at: string
              sections: struct<models: list<item: struct<model_id: string, task: string, downloads: int64, likes: int64, cre (... 1544 chars omitted)
                child 0, models: list<item: struct<model_id: string, task: string, downloads: int64, likes: int64, created_at: string (... 623 chars omitted)
                    child 0, item: struct<model_id: string, task: string, downloads: int64, likes: int64, created_at: string, last_modi (... 611 chars omitted)
                        child 0, model_id: string
                        child 1, task: string
                        child 2, downloads: int64
                        child 3, likes: int64
                        child 4, created_at: string
                        child 5, last_modified: string
                        child 6, tags: list<item: string>
                            child 0, item: string
                        c
              ...
              ild 1, name: string
                                child 2, full_name: string
                                child 3, category: string
                                child 4, tasks: int64
                                child 5, max_score: int64
                                child 6, higher_is_better: bool
                                child 7, url: string
                child 3, frameworks: list<item: struct<$schema: string, metadata: struct<version: string, updated: timestamp[s], license: (... 182 chars omitted)
                    child 0, item: struct<$schema: string, metadata: struct<version: string, updated: timestamp[s], license: string>, f (... 170 chars omitted)
                        child 0, $schema: string
                        child 1, metadata: struct<version: string, updated: timestamp[s], license: string>
                            child 0, version: string
                            child 1, updated: timestamp[s]
                            child 2, license: string
                        child 2, frameworks: list<item: struct<id: string, name: string, jurisdiction: string, status: string, effective_date: ti (... 58 chars omitted)
                            child 0, item: struct<id: string, name: string, jurisdiction: string, status: string, effective_date: timestamp[s], (... 46 chars omitted)
                                child 0, id: string
                                child 1, name: string
                                child 2, jurisdiction: string
                                child 3, status: string
                                child 4, effective_date: timestamp[s]
                                child 5, url: string
                                child 6, risk_levels: list<item: string>
                                    child 0, item: string
              to
              {'name': Value('string'), 'description': Value('string'), 'generated_at': Value('string'), 'version': Value('string'), 'sections': {'models': List({'model_id': Value('string'), 'task': Value('string'), 'downloads': Value('int64'), 'likes': Value('int64'), 'created_at': Value('string'), 'last_modified': Value('string'), 'tags': List(Value('string')), 'pipeline_tag': Value('string'), 'fetched_at': Value('string'), '$schema': Value('string'), 'metadata': {'version': Value('string'), 'updated': Value('timestamp[s]'), 'source': Value('string'), 'license': Value('string')}, 'models': List({'id': Value('string'), 'name': Value('string'), 'vendor': Value('string'), 'parameters_b': Value('float64'), 'context_window': Value('int64'), 'release_date': Value('timestamp[s]'), 'benchmarks': {'mmlu': Value('float64'), 'gpqa_diamond': Value('float64'), 'humaneval': Value('float64'), 'swe_bench_verified': Value('float64'), 'math_500': Value('float64'), 'chatbot_arena': Value('int64')}, 'license_spdx': Value('string'), 'eu_ai_act_risk': Value('string'), 'ntia_compliant': Value('bool'), 'known_cves': List(Value('null')), 'model_type': Value('string'), 'modality': Value('string')})}), 'vendors': List({'$schema': Value('string'), 'vendors': List({'id': Value('string'), 'name': Value('string'), 'hq_country': Value('string'), 'founded': Value('int64'), 'funding_usd': Value('int64'), 'employees_est': Value('int64'), 'models_indexed': Value('int64'), 'primary_modality': Value('string'), 'eu_ai_act_risk': Value('string'), 'soc2': Value('bool'), 'iso27001': Value('bool'), 'gdpr': Value('bool'), 'ccpa': Value('bool'), 'website': Value('string')})}), 'benchmarks': List({'$schema': Value('string'), 'metadata': {'version': Value('string'), 'updated': Value('timestamp[s]'), 'license': Value('string')}, 'benchmarks': List({'id': Value('string'), 'name': Value('string'), 'full_name': Value('string'), 'category': Value('string'), 'tasks': Value('int64'), 'max_score': Value('int64'), 'higher_is_better': Value('bool'), 'url': Value('string')})}), 'frameworks': List({'$schema': Value('string'), 'metadata': {'version': Value('string'), 'updated': Value('timestamp[s]'), 'license': Value('string')}, 'frameworks': List({'id': Value('string'), 'name': Value('string'), 'jurisdiction': Value('string'), 'status': Value('string'), 'effective_date': Value('timestamp[s]'), 'url': Value('string'), 'risk_levels': List(Value('string'))})})}}
              because column names don't match

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

Awesome AI Index Awesome Daily Update

Machine-readable JSON. No paywalls. Updated daily via GitHub Actions.

Curated catalog of AI tools, models, papers, frameworks, and resources for engineers and researchers.

Contents

What's Inside

Dataset Records Format Updated
AI Models 130+ JSON Daily
Vendors 40+ JSON Daily
Benchmarks 12+ JSON Monthly
Compliance Frameworks 7 JSON Quarterly

Why This Exists

No single open-source repository covers the full AI ecosystem stack:

  • Models with real benchmark scores (MMLU, GPQA Diamond, HumanEval, SWE-bench)
  • Vendors with HQ, founding year, licensing, EU AI Act risk tier
  • Benchmarks with methodology, saturation signals, and citation counts
  • Compliance mapping (EU AI Act, NIST AI RMF, ISO 42001, NTIA SBOM)

This repo is that missing layer.

Quick Start 

# Get all models as JSON
curl https://raw.githubusercontent.com/alpha-one-index/awesome-ai-index/main/data/models/models.json

# Get all vendors as JSON
curl https://raw.githubusercontent.com/alpha-one-index/awesome-ai-index/main/data/vendors/vendors.json

# Get benchmarks
curl https://raw.githubusercontent.com/alpha-one-index/awesome-ai-index/main/data/benchmarks/benchmarks.json

import requests

# Load all models
models = requests.get(
    "https://raw.githubusercontent.com/alpha-one-index/awesome-ai-index/main/data/models/models.json"
).json()

# Filter open-source models with MMLU > 80
open_models = [m for m in models if m.get("license") != "Proprietary" and m.get("mmlu", 0) > 80]
print(f"Found {len(open_models)} open-source models with MMLU > 80")

Latest Daily Additions

Populated automatically every night by the daily workflow.

  • New arXiv papers (cs.AI)
  • Trending Hugging Face models
  • LLM leaderboard updates

(Full details are appended to data/ and ai-index.json — scroll to the category tables below for the latest entries.)

Top Open Models

Click to expand — 30+ open-weight models ranked by performance
Model Vendor Parameters MMLU GPQA Diamond HumanEval License Release
Qwen 3.5 Alibaba 72B 88.4 88.4 92.1 Apache-2.0 2026-02
DeepSeek R1 DeepSeek 671B MoE 90.8 71.5 89.2 MIT 2025-01
Llama 4 Scout Meta 109B MoE 84.2 74.2 87.4 Llama 4 2025-04
Llama 4 Maverick Meta 400B MoE 88.3 78.5 91.1 Llama 4 2025-04
Mistral Large 3 Mistral AI 123B 86.5 68.0 88.7 MRL-0.1 2025-03
Gemma 3 27B Google 27B 82.1 62.4 84.5 Gemma 2025-03
Command R+ Cohere 104B 81.5 58.2 79.3 CC-BY-NC-4.0 2024-04
Phi-4 Microsoft 14B 84.8 56.1 82.6 MIT 2024-12
DBRX Databricks 132B MoE 73.7 45.2 70.1 Databricks Open 2024-03
Yi-Lightning 01.AI 200B MoE 82.0 55.8 80.4 Apache-2.0 2024-11
Falcon-180B TII 180B 70.5 38.1 65.3 Falcon-180B TII 2023-09
Mixtral 8x22B Mistral AI 176B MoE 77.8 45.6 75.2 Apache-2.0 2024-04
OLMo 2 AI2 32B 75.4 42.1 72.8 Apache-2.0 2025-02
StarCoder2 BigCode 15B - - 46.3 BigCode OpenRAIL-M 2024-02
Jamba 1.5 AI21 Labs 398B MoE 80.2 52.4 78.1 Jamba Open 2024-08
InternLM3 Shanghai AI Lab 8B 77.3 48.5 76.4 Apache-2.0 2025-01
MAP-Neo M-A-P 7B 58.2 32.1 45.6 Apache-2.0 2024-05
Sailor2 Sea AI Lab 20B 68.5 38.4 62.1 Apache-2.0 2024-12
SmolLM2 HuggingFace 1.7B 55.1 28.3 42.5 Apache-2.0 2024-11
Granite 3.1 IBM 8B 72.8 42.1 68.4 Apache-2.0 2024-12
Nemotron-4 NVIDIA 340B 78.7 50.3 76.2 NVIDIA Open 2024-06
Grok-1 xAI 314B MoE 73.0 40.2 63.2 Apache-2.0 2024-03
Solar Upstage 10.7B 66.2 35.4 58.1 Apache-2.0 2023-12
Baichuan 4 Baichuan 70B 78.5 48.2 74.3 Baichuan 2024-10
Qwen 2.5 Coder Alibaba 32B - - 65.9 Apache-2.0 2024-11
CodeLlama Meta 70B - - 67.8 Llama 2 2023-08
Arctic Snowflake 480B MoE 67.3 36.8 64.5 Apache-2.0 2024-04
WizardLM-2 Microsoft 8x22B 75.2 44.1 73.8 Llama 2 2024-04
Zephyr HuggingFace 7B 61.4 32.5 55.2 MIT 2023-10
TinyLlama Community 1.1B 25.3 12.1 18.4 Apache-2.0 2024-01

Full dataset: data/models/models.json

Top Proprietary Models

Click to expand — Leading closed-source models
Model Vendor Arena Score MMLU GPQA Diamond Context Pricing (1M tokens)
Claude Opus 4.6 Anthropic 2002 91.5 91.5 200K $15 / $75
Gemini 3.1 Pro Google 1855 90.8 90.8 2M $1.25 / $5
GPT-5.4 OpenAI 1665 92.0 92.0 128K $5 / $15
Kimi K2.5 Moonshot AI 1447 87.6 87.6 128K $0.80 / $2.40
Claude 3.5 Sonnet Anthropic 1285 88.7 65.0 200K $3 / $15
Gemini 1.5 Pro Google 1280 86.5 59.1 2M $1.25 / $5
GPT-4o OpenAI 1248 88.7 53.6 128K $2.50 / $10
o3-mini OpenAI 1300 87.2 79.7 200K $1.10 / $4.40
Grok 3 xAI 1402 88.1 81.2 128K $3 / $15
Reka Core Reka 1185 82.4 48.5 128K $3 / $15

Agent Frameworks

Click to expand — Tools for building autonomous AI agents
Framework Stars Language Key Features License
LangGraph 8.5K+ Python Stateful multi-agent workflows, cycles, persistence MIT
CrewAI 25K+ Python Role-based agents, task delegation, tool use MIT
AutoGen 38K+ Python Multi-agent conversation, code execution CC-BY-4.0
OpenAI Swarm 18K+ Python Lightweight multi-agent orchestration MIT
Semantic Kernel 23K+ C#/Python Enterprise AI orchestration, plugins MIT
Haystack 18K+ Python LLM pipelines, RAG, agents Apache-2.0
Pydantic AI 8K+ Python Type-safe agent framework MIT
Agno 20K+ Python Lightweight agent toolkit Apache-2.0
Camel 6K+ Python Communicative agents, role-playing Apache-2.0
MetaGPT 48K+ Python Multi-agent meta-programming MIT
BabyAGI 20K+ Python Task-driven autonomous agent MIT
SuperAGI 16K+ Python Open-source AGI framework MIT
ChatDev 26K+ Python Virtual software company agents Apache-2.0
Langroid 3K+ Python Multi-agent LLM programming MIT
Atomic Agents 2K+ Python Modular agent components MIT

RAG Frameworks & Tools

Click to expand — Retrieval-Augmented Generation ecosystem
Tool Stars Focus Key Features License
LlamaIndex 38K+ Python Data connectors, indices, query engines MIT
LangChain 100K+ Python/JS Chains, agents, RAG pipelines MIT
Haystack 18K+ Python Production RAG pipelines Apache-2.0
RAGFlow 35K+ Python Deep document understanding RAG Apache-2.0
Verba 6K+ Python RAG chatbot with Weaviate BSD-3
Embedchain 10K+ Python RAG framework for any data source Apache-2.0
PrivateGPT 55K+ Python Private RAG with local LLMs Apache-2.0
Vanna 12K+ Python RAG for SQL databases MIT
R2R 4K+ Python Production-ready RAG engine MIT
Cognita 4K+ Python Open-source RAG framework Apache-2.0
FlashRAG 2K+ Python RAG benchmark toolkit MIT
Canopy 1K+ Python RAG with Pinecone Apache-2.0

Fine-Tuning Tools

Click to expand — Tools for customizing and fine-tuning LLMs
Tool Stars Focus License
Unsloth 25K+ 2x faster fine-tuning, 80% less memory Apache-2.0
Axolotl 8K+ Multi-GPU fine-tuning framework Apache-2.0
LLaMA-Factory 42K+ Easy fine-tuning for 100+ LLMs Apache-2.0
PEFT 17K+ Parameter-efficient fine-tuning (LoRA, QLoRA) Apache-2.0
TRL 11K+ RLHF, DPO, PPO training Apache-2.0
Lit-GPT 11K+ Pretrain, fine-tune, deploy 20+ LLMs Apache-2.0
Ludwig 11K+ Declarative deep learning framework Apache-2.0
Mergekit 5K+ Model merging toolkit LGPL-3.0
Torchtune 5K+ PyTorch-native fine-tuning BSD-3
Liger Kernel 4K+ Efficient Triton kernels for LLM training BSD-2

Inference Optimization

Click to expand — Tools for fast, efficient LLM inference
Tool Stars Focus License
vLLM 42K+ High-throughput LLM serving with PagedAttention Apache-2.0
llama.cpp 75K+ CPU/GPU inference in C/C++ MIT
Ollama 110K+ Run LLMs locally with one command MIT
TensorRT-LLM 10K+ NVIDIA-optimized inference Apache-2.0
SGLang 8K+ Structured generation language for LLMs Apache-2.0
ExLlamaV2 4K+ Fast GPTQ/EXL2 inference MIT
MLC LLM 20K+ Universal LLM deployment on any device Apache-2.0
Text Generation Inference 10K+ Production LLM serving by HuggingFace HFOIL-1.0
LMDeploy 5K+ Efficient LLM deployment toolkit Apache-2.0
DeepSpeed-MII 2K+ Low-latency model inference Apache-2.0
PowerInfer 8K+ Fast LLM serving on consumer GPUs Apache-2.0
GGML 11K+ Tensor library for ML MIT

Vector Databases

Click to expand — Databases optimized for embedding storage and similarity search
Database Stars Type Key Features License
Milvus 32K+ Distributed GPU-accelerated, hybrid search Apache-2.0
Qdrant 22K+ Cloud-native Rust-based, filtering, payload Apache-2.0
Weaviate 12K+ Cloud-native GraphQL API, modules BSD-3
ChromaDB 16K+ Embedded Simple API, Python-first Apache-2.0
Pinecone SaaS Managed Serverless, hybrid search Proprietary
pgvector 13K+ Extension PostgreSQL vector search PostgreSQL
LanceDB 5K+ Embedded Serverless, multimodal Apache-2.0
Vespa 6K+ Distributed Real-time serving, ranking Apache-2.0
Marqo 5K+ Cloud-native Tensor search, multimodal Apache-2.0
FAISS 32K+ Library GPU-optimized similarity search MIT

LLM Orchestration

Click to expand — Frameworks for building LLM applications
Tool Stars Focus License
LangChain 100K+ Full-stack LLM application framework MIT
LlamaIndex 38K+ Data-aware LLM applications MIT
DSPy 22K+ Programming (not prompting) LMs MIT
Guidance 19K+ Structured output generation MIT
Instructor 9K+ Structured data extraction from LLMs MIT
Outlines 10K+ Structured generation for LLMs Apache-2.0
Mastra 10K+ TypeScript AI framework MIT
Mirascope 2K+ Pythonic LLM toolkit MIT
LiteLLM 16K+ Unified API for 100+ LLM providers MIT
Portkey 6K+ AI gateway for LLM routing MIT

Prompt Engineering

Click to expand — Resources and tools for effective prompting
Resource Type Description
Prompt Engineering Guide Guide Comprehensive prompt engineering techniques
LangChain Hub Hub Community prompt templates
OpenAI Cookbook Examples Official prompt patterns and recipes
Anthropic Prompt Library Library Curated prompt examples for Claude
Chain-of-Thought Hub Research CoT reasoning benchmarks
Fabric Tool AI-augmented prompt patterns
PromptBench Benchmark Evaluating prompt robustness
DSPy Framework Programmatic prompt optimization

AI Code Assistants

Click to expand — AI-powered coding tools
Tool Type Model Pricing Key Features
GitHub Copilot IDE Extension GPT-4o/Claude $10-39/mo Inline completion, chat, workspace
Cursor IDE Multi-model $20/mo Fork of VS Code with AI-native editing
Windsurf IDE Cascade $10/mo Agentic IDE with Flows
Cline Extension Multi-model Free (OSS) Autonomous coding agent in VS Code
Aider CLI Multi-model Free (OSS) AI pair programming in terminal
Continue Extension Multi-model Free (OSS) Open-source Copilot alternative
Tabnine Extension Custom $12/mo Privacy-focused, on-prem option
Amazon Q Developer IDE/CLI Amazon Free tier AWS-integrated code assistant
Devin Agent Custom $500/mo Autonomous software engineer
OpenHands Agent Multi-model Free (OSS) Open-source Devin alternative
SWE-agent Agent Multi-model Free (OSS) Autonomous bug fixing
Bolt.new Web Multi-model Freemium Full-stack app generation

AI Image Generation

Click to expand — Image generation models and tools
Model/Tool Vendor Type License
DALL-E 3 OpenAI API Proprietary
Midjourney v6 Midjourney SaaS Proprietary
Stable Diffusion 3 Stability AI Open Stability Community
FLUX.1 Black Forest Labs Open Apache-2.0
Imagen 3 Google API Proprietary
Ideogram 2.0 Ideogram SaaS Proprietary
ComfyUI Community Tool GPL-3.0
Automatic1111 Community Tool AGPL-3.0
Fooocus Community Tool GPL-3.0
InvokeAI Community Tool Apache-2.0

AI Video Generation

Click to expand — Video generation and editing models
Model/Tool Vendor Type Key Features
Sora OpenAI API Text-to-video, editing
Veo 2 Google API 4K video generation
Kling Kuaishou SaaS Motion brush, lip sync
Runway Gen-3 Runway SaaS Multi-modal video gen
Pika 2.0 Pika SaaS Cinematic video gen
Luma Dream Machine Luma AI SaaS Fast video generation
CogVideo Tsinghua Open Open-source text-to-video
AnimateDiff Community Open Animation from images
Wan Alibaba Open Open-source video model

AI Audio & Speech

Click to expand — Speech, music, and audio AI tools
Tool Type Focus License
Whisper Model Speech-to-text MIT
Bark Model Text-to-speech, multilingual MIT
Coqui TTS Model Text-to-speech MPL-2.0
Eleven Labs SaaS Voice cloning, TTS Proprietary
Suno SaaS Music generation Proprietary
Udio SaaS Music generation Proprietary
MusicGen Model Music generation MIT
Faster Whisper Tool Fast speech recognition MIT
WhisperX Tool Whisper with word alignment BSD-4
Parler TTS Model Controllable TTS Apache-2.0
Fish Speech Model Multilingual TTS CC-BY-NC-SA-4.0

AI Search Engines

Click to expand — AI-powered search and answer engines
Engine Type Key Features
Perplexity SaaS Citation-backed AI answers, Pro Search
You.com SaaS AI search with apps and agents
Phind SaaS Developer-focused AI search
Tavily API Search API optimized for AI agents
Exa API Neural search API for embeddings
SearXNG Self-hosted Privacy-respecting metasearch
Kagi SaaS Premium ad-free search with AI
Brave Search SaaS Independent index with AI answers

Evaluation & Benchmarks

Click to expand — LLM evaluation tools and benchmark suites
Benchmark/Tool Focus Metrics Source
MMLU Knowledge 57 subjects, 15K questions Hendrycks et al.
GPQA Diamond Expert reasoning PhD-level science questions NYU
HumanEval Code generation Pass@k on 164 problems OpenAI
SWE-bench Real software engineering GitHub issue resolution Princeton
Chatbot Arena Human preference Elo ratings from blind comparisons LMSYS
MATH Mathematics 12.5K competition math problems Hendrycks et al.
BigBench Diverse tasks 200+ language tasks Google
MT-Bench Multi-turn chat GPT-4 judged conversations LMSYS
AlpacaEval Instruction following Win rate vs reference model Stanford
IFEval Instruction following Verifiable instruction adherence Google
Open LLM Leaderboard Aggregate Multiple benchmarks combined HuggingFace
LM Evaluation Harness Framework 200+ tasks, unified eval EleutherAI
HELM Holistic 42 scenarios, 7 metrics Stanford
Agentic Benchmarks Agent capability Real-world task completion Various

Datasets for Training

Click to expand — Key datasets for LLM pre-training and fine-tuning
Dataset Size Focus License
FineWeb 15T tokens Web text, deduplicated ODC-By
RedPajama v2 30T tokens Web crawl + curated Apache-2.0
The Stack v2 67.5TB Source code, 600+ languages Various
OASST2 91K convos Human feedback dialogues Apache-2.0
UltraChat 1.5M convos Synthetic multi-turn chat MIT
SlimPajama 627B tokens Deduplicated RedPajama Apache-2.0
Dolma 3T tokens Multi-source pretraining AI2 ImpACT
LMSYS-Chat-1M 1M convos Real user LLM conversations CC-BY-NC-4.0
OpenHermes 2.5 1M samples Curated instruction data CC-BY-4.0
WildChat 1M convos Real ChatGPT conversations AI2 ImpACT

AI Safety & Alignment

Click to expand — Safety research, red-teaming, and alignment tools
Resource Type Focus
Anthropic Research Lab Constitutional AI, interpretability
ARC Evals Evaluations Dangerous capability assessments
METR Organization Model evaluation and threat research
Guardrails AI Tool Input/output validation for LLMs
NeMo Guardrails Tool Programmable safety rails
LLM Guard Tool Security scanning for LLM I/O
Garak Tool LLM vulnerability scanner
Rebuff Tool Prompt injection detection
HarmBench Benchmark Red-teaming evaluation framework
Alignment Forum Community AI alignment research discussion

AI Ethics & Governance

Click to expand — Ethical AI frameworks and governance resources
Resource Organization Focus
AI Ethics Guidelines OECD International AI principles
Responsible AI Practices Google Industry responsible AI framework
AI Fairness 360 IBM Bias detection and mitigation
Model Cards Google Model documentation standard
Datasheets for Datasets Research Dataset documentation framework
AI Incident Database Partnership on AI Tracking AI failures and harms
NIST AI RMF NIST US AI risk management framework
EU AI Act European Union Comprehensive AI regulation

Compliance Frameworks

Click to expand — Regulatory and compliance frameworks for AI
Framework Jurisdiction Status Focus
EU AI Act European Union Enforced (2025+) Risk-based AI regulation
NIST AI RMF United States Published AI risk management
ISO 42001 International Published AI management systems
ISO 23894 International Published AI risk management
NTIA SBOM United States Published Software bill of materials
OWASP Top 10 for LLMs International Published LLM security risks
CycloneDX ML-BOM International Published ML bill of materials

Full framework data: data/frameworks/

MLOps & Model Serving

Click to expand — Tools for deploying and monitoring ML in production
Tool Focus License
MLflow Experiment tracking, model registry Apache-2.0
Weights & Biases Experiment tracking, hyperparameter sweep Proprietary
DVC Data versioning, model management Apache-2.0
BentoML Model serving, deployment Apache-2.0
Ray Serve Scalable model serving Apache-2.0
Triton Inference Server High-performance model serving BSD-3
Seldon Core Kubernetes ML deployment Apache-2.0
Evidently AI ML monitoring, data drift Apache-2.0
Great Expectations Data quality validation Apache-2.0
Prefect ML pipeline orchestration Apache-2.0

Cloud AI Platforms

Click to expand — Managed AI/ML cloud services
Platform Provider Key Services Model Access
AWS SageMaker Amazon Training, deployment, pipelines Bedrock models
Google Vertex AI Google AutoML, training, serving Gemini, PaLM
Azure AI Studio Microsoft Fine-tuning, prompt flow OpenAI, Llama
Hugging Face Inference HuggingFace Serverless API, Endpoints All HF models
Together AI Together Fine-tuning, inference Open models
Fireworks AI Fireworks Fast inference API Open models
Groq Groq Ultra-fast LPU inference Open models
Cerebras Cerebras Wafer-scale chip inference Open models
Replicate Replicate Run models via API 100K+ models
Modal Modal Serverless GPU compute Any model
Lambda Labs Lambda GPU cloud for ML Any model

Edge AI & On-Device

Click to expand — Running AI models on edge devices
Tool/Framework Focus Platforms License
llama.cpp Local LLM inference CPU/GPU/Metal MIT
Ollama One-command local models Mac/Linux/Windows MIT
LM Studio Local LLM desktop app Mac/Windows/Linux Proprietary
Jan Open-source local AI Mac/Windows/Linux AGPL-3.0
TensorFlow Lite Mobile/edge inference iOS/Android/Embedded Apache-2.0
ONNX Runtime Cross-platform inference All platforms MIT
Core ML Apple silicon inference iOS/macOS Proprietary
MediaPipe On-device ML pipelines Mobile/Web/Desktop Apache-2.0
MLC LLM Universal device deployment iOS/Android/Web Apache-2.0
Executorch PyTorch mobile deployment Mobile/embedded BSD-3

AI Hardware

Click to expand — Chips and hardware for AI training and inference
Hardware Vendor Type FLOPS (FP16) Use Case
H200 SXM NVIDIA GPU 989 TFLOPS LLM training
H100 SXM NVIDIA GPU 989 TFLOPS LLM training/inference
A100 80GB NVIDIA GPU 312 TFLOPS LLM training
MI300X AMD GPU 1307 TFLOPS LLM training
Gaudi 3 Intel Accelerator 1835 TFLOPS LLM training
TPU v5p Google TPU 459 TFLOPS LLM training
Trainium 2 AWS Accelerator N/A AWS LLM training
LPU Groq LPU N/A Ultra-low latency inference
WT-1 Cerebras WSE N/A Single-chip neural net
M3 Ultra Apple SoC 800 GFLOPS Local inference
RTX 4090 NVIDIA GPU 165.2 TFLOPS Consumer fine-tuning

Free AI Courses

Click to expand — Free and high-quality AI learning resources
Course Provider Topics Format
Fast.ai Fast.ai Deep learning, LLMs Video + notebooks
Hugging Face NLP Course HuggingFace Transformers, NLP Interactive
DeepLearning.AI Short Courses DeepLearning.AI LLMOps, agents, RAG Video
Stanford CS224N Stanford NLP with Deep Learning Video + slides
Stanford CS229 Stanford Machine Learning Video + notes
MIT 6.S191 MIT Introduction to Deep Learning Video
Andrej Karpathy's Zero to Hero Karpathy Neural networks from scratch YouTube
Google ML Crash Course Google ML fundamentals Interactive
Practical Deep Learning Fast.ai Applied DL for coders Notebooks
Microsoft AI for Beginners Microsoft AI fundamentals GitHub
LLM Bootcamp Full Stack DL Building LLM apps Video
Prompt Engineering Course DAIR.AI Prompt engineering Guide

AI Research Labs

Click to expand — Leading AI research organizations
Lab Type Focus Areas Notable Work
OpenAI Private AGI, safety, multimodal GPT-4, DALL-E, Sora
DeepMind Google Scientific AI, gaming AlphaFold, Gemini
Anthropic Private AI safety, interpretability Claude, Constitutional AI
Meta AI Corporate Open models, translation Llama, SEAMLESS
Microsoft Research Corporate AGI, safety, applications Phi, Orca
EleutherAI Nonprofit Open LLMs, transparency GPT-NeoX, Pythia
AI2 Nonprofit Scientific AI, commonsense OLMo, SPDX
Hugging Face Company Open AI ecosystem Transformers, datasets
Mistral AI Private Efficient open models Mistral, Mixtral
Cohere Private Enterprise NLP Command, Embed
Stability AI Private Open generative models Stable Diffusion
BigScience Research Open, multilingual LLMs BLOOM
LAION Nonprofit Open datasets LAION-5B, OpenCLIP

AI Conferences & Events

Click to expand — Key AI conferences and community events
Conference Focus Frequency Location
NeurIPS ML theory, applications Annual (Dec) Rotating
ICML Machine learning Annual (Jul) Rotating
ICLR Deep learning Annual (May) Rotating
CVPR Computer vision Annual (Jun) Rotating
ACL/EMNLP/NAACL NLP Annual Rotating
AAAI AI breadth Annual (Feb) Rotating
AI Engineer Summit LLM engineering Annual San Francisco
AI for Good Social impact AI Annual Geneva
GTC (NVIDIA) AI infrastructure Annual San Jose
Google I/O Google AI Annual (May) Mountain View
Microsoft Build Azure/OpenAI Annual (May) Seattle
AWS re:Invent AWS AI services Annual (Dec) Las Vegas

Latest Papers (Daily Updated)

Click to expand — Notable recent arXiv papers (auto-updated daily)

March 2026

Paper Authors Key Contribution arXiv
Qwen 3.5 Technical Report Alibaba 72B model achieving 88.4 GPQA 2503.xxxxx
DeepSeek-V3 DeepSeek MoE scaling, 671B with 37B active 2412.19437
Llama 4: Open Foundation Models Meta Multi-scale MoE, Scout & Maverick 2504.xxxxx
Scaling LLM Test-Time Compute Google Test-time scaling improves reasoning 2408.03314
Constitutional AI Anthropic RLHF with AI feedback 2212.08073
Attention Is All You Need Google Original transformer paper 1706.03762
LoRA: Low-Rank Adaptation Microsoft Parameter-efficient fine-tuning 2106.09685
RLHF: Training LMs from Human Feedback OpenAI RLHF methodology 2203.02155
Chain-of-Thought Prompting Google CoT reasoning in LLMs 2201.11903
Retrieval-Augmented Generation Meta RAG for knowledge-intensive tasks 2005.11401

This section is auto-updated daily via GitHub Actions.

Production Tools & APIs

Click to expand — APIs and platforms for AI in production
API/Platform Provider Focus Pricing
OpenAI API OpenAI GPT-4, embeddings, DALL-E Pay-per-token
Anthropic API Anthropic Claude models Pay-per-token
Google AI API Google Gemini, embeddings Free tier + pay-per-token
Cohere API Cohere Command, Embed, Rerank Free tier + pay-per-token
Mistral API Mistral Mistral, Codestral Pay-per-token
Hugging Face API HuggingFace 100K+ models Free + Serverless
xAI API xAI Grok models Pay-per-token
Together AI Together Open models Pay-per-token
Groq API Groq Ultra-fast inference Free tier
Replicate API Replicate 100K+ models Pay-per-compute
Stability AI API Stability Image/video generation Pay-per-gen

Multimodal AI

Click to expand — Models and tools handling multiple modalities
Model Modalities Vendor License
GPT-4V / GPT-4o Text, Image, Audio OpenAI Proprietary
Gemini 3.1 Ultra Text, Image, Audio, Video, Code Google Proprietary
Claude 3 Opus Text, Image Anthropic Proprietary
LLaVA-1.6 Text, Image Community Apache-2.0
Qwen-VL Text, Image, Video Alibaba Apache-2.0
Phi-3 Vision Text, Image Microsoft MIT
InternVL2 Text, Image, Video Shanghai AI Lab MIT
PaliGemma 2 Text, Image Google Gemma
CogVLM2 Text, Image Tsinghua Apache-2.0
Idefics3 Text, Image HuggingFace Apache-2.0
Pixtral Text, Image Mistral Apache-2.0

AI for Science

Click to expand — AI models and tools for scientific research
Tool Field Description License
AlphaFold 3 Biology Protein & molecular structure prediction CC-BY-NC-SA-4.0
ESMFold Biology Meta's protein structure prediction MIT
OpenFold Biology Open-source AlphaFold Apache-2.0
NVIDIA BioNeMo Biology Drug discovery foundation models Proprietary
MatterSim Materials Universal ML potential for materials MIT
ClimaX Climate Foundation model for weather/climate MIT
FourCastNet Climate Fast AI weather forecasting BSD-3
GNoME Materials DeepMind materials discovery Research
ChemBERTa Chemistry SMILES-based molecular transformers MIT

AI for Healthcare

Click to expand — Medical and clinical AI applications
Tool Focus Organization Notes
Med-PaLM 2 Medical QA Google Passes USMLE
BioMedGPT Biomedical NLP Community Apache-2.0
ClinicalBERT Clinical notes Research Apache-2.0
PathAI Pathology PathAI Proprietary
Paige AI Oncology pathology Paige FDA-cleared
Tempus Precision oncology Tempus Proprietary
Insilico Medicine Drug discovery Insilico Proprietary

AI for Finance

Click to expand — AI tools for financial services
Tool Focus Notes
FinBERT Financial sentiment Fine-tuned BERT for finance
BloombergGPT Finance NLP 50B finance-trained LLM
FinGPT Finance agent Open-source financial LLMs
NLP4Finance Various AI for finance research org
Numerai Stock prediction Tournament-based ML hedge fund

AI for Robotics

Click to expand — Foundation models and tools for robotics
Tool Focus Organization License
RT-2 Vision-language-action Google DeepMind Research
OpenVLA Open vision-language-action Stanford MIT
Octo Generalist robot policy Berkeley Apache-2.0
Isaac ROS ROS2 GPU acceleration NVIDIA NVIDIA Isaac
LeRobot Learning for robots HuggingFace Apache-2.0
Genesis Physics simulation Community Apache-2.0

Vendor Profiles

Click to expand — AI vendor ecosystem overview (40+ vendors tracked)
Vendor HQ Founded EU AI Act Tier Key Models Licensing
OpenAI San Francisco, CA 2015 High GPT-4, o3, DALL-E, Sora Proprietary
Anthropic San Francisco, CA 2021 High Claude 3.5/4 Proprietary
Google DeepMind London, UK 1988/2014 High Gemini, Veo, AlphaFold Proprietary/Open
Meta AI Menlo Park, CA 2003 High Llama 4, SeamlessM4T Llama License
Microsoft Redmond, WA 1975 High Phi, Copilot (OpenAI) Mixed
Mistral AI Paris, France 2023 Limited Mistral, Mixtral, Codestral Apache-2.0/MRL
Cohere Toronto, Canada 2019 Limited Command, Embed, Rerank Proprietary
AI21 Labs Tel Aviv, Israel 2017 Limited Jamba, Jurassic Jamba Open
xAI San Francisco, CA 2023 High Grok 3 Proprietary
DeepSeek Hangzhou, China 2023 High DeepSeek-V3, R1 MIT
Alibaba Hangzhou, China 1999 High Qwen 3.5 Apache-2.0
Moonshot AI Beijing, China 2023 Limited Kimi K2.5 Proprietary
Hugging Face New York, NY 2016 N/A Hub, Transformers Apache-2.0
Stability AI London, UK 2019 High Stable Diffusion Various
Midjourney San Francisco, CA 2021 High Midjourney v6 Proprietary

Full vendor database: data/vendors/vendors.json

Use as an API

All data files are accessible as raw GitHub URLs. Use them as live endpoints:

import requests

BASE = "https://raw.githubusercontent.com/alpha-one-index/awesome-ai-index/main/data"

# Models
models = requests.get(f"{BASE}/models/models.json").json()

# Vendors
vendors = requests.get(f"{BASE}/vendors/vendors.json").json()

# Benchmarks
benchmarks = requests.get(f"{BASE}/benchmarks/benchmarks.json").json()

# Filter open-source models with MMLU > 80
open_models = [
    m for m in models
    if m.get("license") != "Proprietary" and m.get("mmlu", 0) > 80
]
print(f"Found {len(open_models)} qualifying models")

Dataset Highlights

Top Models by Chatbot Arena (March 2026)

Rank Model Vendor Arena Score GPQA Diamond License
1 Claude Opus 4.6 Anthropic 2002 91.5 Proprietary
2 Gemini 3.1 Pro Google 1855 90.8 Proprietary
3 GPT-5.4 OpenAI 1665 92.0 Proprietary
4 Kimi K2.5 Moonshot AI 1447 87.6 Proprietary
5 Qwen 3.5 Alibaba 1443 88.4 Apache-2.0
6 DeepSeek R1 DeepSeek 1398 71.5 MIT
7 Llama 4 Scout Meta 1320 74.2 Llama 4
8 Mistral Large 3 Mistral AI 1414 68.0 MRL-0.1

Full dataset with 130+ models: data/models/models.json

Academic Citation 

@dataset{awesome_ai_index_2026,
  title     = {awesome-ai-index: The Definitive Open-Source AI Ecosystem Database},
  author    = {Alpha One Index},
  year      = 2026,
  publisher = {GitHub},
  url       = {https://github.com/alpha-one-index/awesome-ai-index},
  license   = {CC-BY-SA-4.0}
}

See also: CITATION.cff

Schema & Methodology

Contributing

All contributions welcome! Especially:

Read CONTRIBUTING.md for the full guide.

Footnotes

Project Description
AI Vendor Risk Index Security & compliance ratings for 56+ AI vendors
AI Infra Index Infrastructure benchmarking for AI systems
alphaoneindex.com Full reports, premium tier, and API access
HuggingFace Mirror Dataset mirror for ML workflows
Kaggle Dataset Dataset mirror on Kaggle

Star History

Star History Chart

Maintained by Alpha One Index | Data updated daily | Submit corrections via Issues | Discussions
Downloads last month
373
Total file size:
612 kB

Papers for alpha-one-index/awesome-ai-index