AI-PORTAL
memory

Model Leaderboard

Top AI models ranked by benchmark · manually updated

14 models
ModelTierContextunfold_moreSWE-benchunfold_moreMMLUarrow_downwardInput $/1Munfold_moreOutput $/1M
Llama 4 MaverickMeta
balanced1000K-85.5%OSSOSS
Claude Fable 5Anthropic
flagship1000K80.3%-$10.00$50.00
Claude Opus 4.8Anthropic
flagship1000K88.6%-$5.00$25.00
GPT-5.5OpenAI
flagship1049K--$5.00$30.00
Gemini 3.1 ProGoogle
flagship1000K80.6%-$2.00$12.00
Grok 4.3xAI
flagship1000K74.9%-$1.25$2.50
DeepSeek V4 ProDeepSeek
flagship1000K80.6%-$1.74$3.48
Gemini 2.5 Pro Deep ThinkGoogle
reasoning1000K--$3.50$10.50
Claude Sonnet 4.6Anthropic
balanced200K--$3.00$15.00
Qwen 3.5 235BAlibaba
balanced262K--OSSOSS
Claude Haiku 4.5Anthropic
fast200K--$0.80$4.00
Gemini 2.5 FlashGoogle
fast1000K--$0.15$0.60
DeepSeek V4 FlashDeepSeek
fast1000K--$0.28$0.42
Llama 4 ScoutMeta
fast10000K--OSSOSS

SWE-bench = real GitHub coding tasks (higher = better). MMLU = general knowledge. Prices from provider APIs. OSS models: self-host or use via API providers. Last updated Jun 2026.

build

Tools Directory

Frameworks, platforms, and infra for GenAI development

LangChainframework

Most popular framework for building LLM-powered applications, chains, and agents.

terminalOSSstar95kopen_in_new
LlamaIndexframework

Data framework for LLM applications — RAG, structured extraction, and query pipelines.

terminalOSSstar37kopen_in_new
Vercel AI SDKframework

TypeScript toolkit for building AI-powered UIs with streaming, tool use, and multi-modal support.

terminalOSSstar12kopen_in_new
Haystackframework

Production-ready NLP framework for building custom pipelines with RAG, search, and agents.

terminalOSSstar17kopen_in_new
Semantic Kernelframework

Microsoft's SDK for integrating LLMs into C#, Python, and Java enterprise applications.

terminalOSSstar22kopen_in_new
CrewAImulti-agent

Framework for orchestrating role-playing, autonomous AI agents working as a team.

terminalOSSstar23kopen_in_new
AutoGenmulti-agent

Microsoft's framework for building multi-agent applications where agents converse to solve tasks.

terminalOSSstar34kopen_in_new
LangGraphmulti-agent

Build stateful, multi-actor agent applications as controllable graphs with cycles and branches.

terminalOSSstar8kopen_in_new
Difyplatform

Visual platform for building and operating AI applications with drag-and-drop workflows.

terminalOSSstar45kopen_in_new
Flowiseplatform

Drag-and-drop UI for building LLM flows and AI agents with LangChain components.

terminalOSSstar31kopen_in_new
n8n AIplatform

Workflow automation platform with native AI nodes for building agentic workflows without code.

terminalOSSstar47kopen_in_new
Ollamalocal

Run LLMs locally with a simple CLI. Supports Llama, Mistral, Gemma, Phi and hundreds more.

terminalOSSstar90kopen_in_new
LM Studiolocal

Desktop app to discover, download, and run local LLMs with a ChatGPT-like UI.

open_in_new
vLLMlocal

High-throughput, memory-efficient inference engine for LLMs. The standard for production self-hosting.

terminalOSSstar31kopen_in_new
Model Context Protocolmcp

Anthropic's open standard for connecting AI assistants to data sources and tools.

terminalOSSstar12kopen_in_new
OpenAI Assistantsmcp

Hosted platform for building AI assistants with persistent threads, code interpreter, and file search.

open_in_new
Weights & Biasesmlops

ML experiment tracking, model versioning, and dataset management for the full ML lifecycle.

star9kopen_in_new
LangSmithmlops

Observability and testing platform for LLM applications built with LangChain.

open_in_new
Langfusemlops

Open-source LLM engineering platform for tracing, evaluation, and prompt management.

terminalOSSstar7kopen_in_new
Pineconevector

Managed vector database built for production AI applications with low latency at scale.

open_in_new
Chromavector

Open-source embedding database for AI applications. Simple API, runs embedded or as a server.

terminalOSSstar15kopen_in_new
Qdrantvector

High-performance vector search engine with rich filtering, payload indexing, and Rust core.

terminalOSSstar20kopen_in_new
Weaviatevector

Open-source vector database with built-in ML model integrations and GraphQL API.

terminalOSSstar12kopen_in_new
map

AI Vendor Map

34 vendors · 6 categories · Updated June 2026

psychology

Foundation Models

(8)
cloud

Cloud Platforms

(6)
code

Frameworks

(5)
account_tree

Agentic / Auto

(6)
monitor_heart

MLOps / Observ.

(4)
database

Vector DBs

(5)
schema

AI Architecture Patterns

7 patterns · When to use each · Common pitfalls · June 2026

chat

Basic Prompt Engineering

Low

Simple Q&A, text transformation, classification, summarization. No external data needed. Fastest to ship.

expand_more
database

RAG (Retrieval-Augmented Generation)

Medium

Q&A over company documents, knowledge bases, or any proprietary content. Prevents hallucination on factual queries.

expand_more
build

Tool Use / Function Calling

Medium

LLM needs to interact with external systems: search the web, query a database, call an API, execute code.

expand_more
account_tree

Agentic Loop (ReAct)

High

Multi-step tasks where the next action depends on previous results. The LLM acts as a planner and executor.

expand_more
groups

Multi-Agent System

High

Parallel subtasks, specialization, or peer-review between agents. Scale beyond single-agent context limits.

expand_more
model_training

Fine-tuned Model

High

Consistent domain-specific behavior at scale. Style, format, or domain knowledge the base model lacks. High-volume inference where prompt size matters for cost.

expand_more
security

Guardrails & Safety Layer

Medium

Any production deployment. Add input/output validation to prevent prompt injection, PII leakage, and off-topic responses.

expand_more