/ai-models

name: ai-models
description: Latest AI models reference - Claude, OpenAI, Gemini, Eleven Labs, Replicate
when-to-use: When choosing models, comparing capabilities, or referencing model specs
user-invocable: true
effort: low

AI Models Reference Skill

**Last Updated: December 2025**

Philosophy

**Use the right model for the job.** Bigger isn't always better - match model capabilities to task requirements. Consider cost, latency, and accuracy tradeoffs.

Model Selection Matrix

| Task | Recommended | Why | |------|-------------|-----| | Complex reasoning | Claude Opus 4.5, o3, Gemini 3 Pro | Highest accuracy | | Fast chat/completion | Claude Haiku, GPT-4.1 mini, Gemini Flash | Low latency, cheap | | Code generation | Claude Sonnet 4.5, Codestral, GPT-4.1 | Strong coding | | Vision/images | Claude Sonnet, GPT-4o, Gemini 3 Pro | Multimodal | | Embeddings | text-embedding-3-small, Voyage | Cost-effective | | Voice synthesis | Eleven Labs v3, OpenAI TTS | Natural sounding | | Image generation | FLUX.2, DALL-E 3, SD 3.5 | Different styles |

---

Anthropic (Claude)

Documentation

• **API Docs**: https://docs.anthropic.com
• **Models Overview**: https://docs.anthropic.com/en/docs/about-claude/models/overview
• **Pricing**: https://www.anthropic.com/pricing

Latest Models (December 2025)

const CLAUDE_MODELS = {
  // Flagship - highest capability
  opus: 'claude-opus-4-5-20251101',

  // Balanced - best for most tasks
  sonnet: 'claude-sonnet-4-5-20250929',

  // Previous generation (still excellent)
  opus4: 'claude-opus-4-6',
  sonnet4: 'claude-sonnet-5',

  // Fast & cheap - high volume tasks
  haiku: 'claude-haiku-4-5-20251001',
} as const;

Usage

import Anthropic from '@anthropic-ai/sdk';

const anthropic = new Anthropic({
  apiKey: process.env.ANTHROPIC_API_KEY,
});

const response = await anthropic.messages.create({
  model: 'claude-sonnet-4-5-20250929',
  max_tokens: 1024,
  messages: [
    { role: 'user', content: 'Hello, Claude!' }
  ],
});

Model Selection

claude-opus-4-5-20251101 (Opus 4.5)
├── Best for: Complex analysis, research, nuanced writing
├── Context: 200K tokens
├── Cost: $5/$25 per 1M tokens (input/output)
└── Use when: Accuracy matters most

claude-sonnet-4-5-20250929 (Sonnet 4.5)
├── Best for: Code, general tasks, balanced performance
├── Context: 200K tokens
├── Cost: $3/$15 per 1M tokens
└── Use when: Default choice for most applications

claude-haiku-4-5-20251001 (Haiku 4.5)
├── Best for: Classification, extraction, high-volume
├── Context: 200K tokens
├── Cost: $1/$5 per 1M tokens
└── Use when: Speed and cost matter most

---

OpenAI

Documentation

• **API Docs**: https://platform.openai.com/docs
• **Models**: https://platform.openai.com/docs/models
• **Pricing**: https://openai.com/pricing

Latest Models (December 2025)

const OPENAI_MODELS = {
  // GPT-5 series (latest)
  gpt5: 'gpt-5.2',
  gpt5Mini: 'gpt-5-mini',

  // GPT-4.1 series (recommended for most)
  gpt41: 'gpt-4.1',
  gpt41Mini: 'gpt-4.1-mini',
  gpt41Nano: 'gpt-4.1-nano',

  // Reasoning models (o-series)
  o3: 'o3',
  o3Pro: 'o3-pro',
  o4Mini: 'o4-mini',

  // Legacy but still useful
  gpt4o: 'gpt-4o',           // Still has audio support
  gpt4oMini: 'gpt-4o-mini',

  // Embeddings
  embeddingSmall: 'text-embedding-3-small',
  embeddingLarge: 'text-embedding-3-large',

  // Image generation
  dalle3: 'dall-e-3',
  gptImage: 'gpt-image-1',

  // Audio
  tts: 'tts-1',
  ttsHd: 'tts-1-hd',
  whisper: 'whisper-1',
} as const;

Usage

import OpenAI from 'openai';

const openai = new OpenAI({
  apiKey: process.env.OPENAI_API_KEY,
});

// Chat completion
const response = await openai.chat.completions.create({
  model: 'gpt-4.1',
  messages: [
    { role: 'user', content: 'Hello!' }
  ],
});

// With vision
const visionResponse = await openai.chat.completions.create({
  model: 'gpt-4.1',
  messages: [
    {
      role: 'user',
      content: [
        { type: 'text', text: 'What is in this image?' },
        { type: 'image_url', image_url: { url: 'https://...' } },
      ],
    },
  ],
});

// Embeddings
const embedding = await openai.embeddings.create({
  model: 'text-embedding-3-small',
  input: 'Your text here',
});

Model Selection

o3 / o3-pro
├── Best for: Math, coding, complex multi-step reasoning
├── Context: 200K tokens
├── Cost: Premium pricing
└── Use when: Hardest problems, need chain-of-thought

gpt-4.1
├── Best for: General tasks, coding, instruction following
├── Context: 1M tokens (!)
├── Cost: Lower than GPT-4o
└── Use when: Default choice, replaces GPT-4o

gpt-4.1-mini / gpt-4.1-nano
├── Best for: High-volume, cost-sensitive
├── Context: 1M tokens
├── Cost: Very low
└── Use when: Simple tasks at scale

o4-mini
├── Best for: Fast reasoning at low cost
├── Context: 200K tokens
├── Cost: Budget reasoning
└── Use when: Need reasoning but cost-conscious

---

Google (Gemini)

Documentation

• **API Docs**: https://ai.google.dev/docs
• **Models**: https://ai.google.dev/gemini-api/docs/models/gemini
• **Pricing**: https://ai.google.dev/pricing

Latest Models (December 2025)

const GEMINI_MODELS = {
  // Gemini 3 (Latest)
  gemini3Pro: 'gemini-3-pro-preview',
  gemini3ProImage: 'gemini-3-pro-image-preview',
  gemini3Flash: 'gemini-3-flash-preview',

  // Gemini 2.5 (Stable)
  gemini25Pro: 'gemini-2.5-pro',
  gemini25Flash: 'gemini-2.5-flash',
  gemini25FlashLite: 'gemini-2.5-flash-lite',

  // Specialized
  gemini25FlashTTS: 'gemini-2.5-flash-preview-tts',
  gemini25FlashAudio: 'gemini-2.5-flash-native-audio-preview-12-2025',

  // Previous generation
  gemini2Flash: 'gemini-2.0-flash',
} as const;

Usage

import { GoogleGenerativeAI } from '@google/generative-ai';

const genAI = new GoogleGenerativeAI(process.env.GOOGLE_API_KEY);
const model = genAI.getGenerativeModel({ model: 'gemini-2.5-flash' });

const result