Frontier LLMs

AI Models Hub

Every frontier AI model, benchmarked, priced and compared. Updated as new models ship.

Model vendors

OpenAI · 3Anthropic · 1Google · 1xAI · 1Meta · 1Mistral · 1DeepSeek · 1Alibaba · 1Perplexity · 1

All models

  • GPT-6 Astra

    OpenAI · Released 2026-09-18

    reasoning

    OpenAI's next-gen frontier model with native multi-step reasoning, autonomous coding & web building.

    Context
    256K tokens
    Pricing
    Plus $20/mo · Pro $200/mo · API pay-as-you-go
    Strengths
    • Breakthrough multi-step reasoning
    • Autonomous full-stack code synthesis
    • Native website compilation
    Trade-offs
    • Higher API token costs on pro tier
    Official site →Compare models

  • GPT-4o

    OpenAI · Released 2024-05-13

    chat

    OpenAI's omni-modal flagship — voice, vision and text in one model.

    Context
    128K tokens
    Pricing
    $2.50 / $10 per 1M tokens · Free tier · Plus $20/mo
    Strengths
    • Native multimodal
    • Fastest GPT-4-class
    • Massive ecosystem
    Trade-offs
    • Smaller context than Claude
    • Occasional hallucinations
    Official site →Compare models

  • Claude 3.5 Sonnet

    Anthropic · Released 2024-06-20

    reasoning

    Anthropic's deep-reasoning model — best-in-class coding and long context.

    Context
    200K tokens
    Pricing
    $3 / $15 per 1M tokens · Free tier · Pro $20/mo
    Strengths
    • SOTA on SWE-bench
    • 200K context
    • Best prose quality
    Trade-offs
    • No native image gen
    • Stricter refusals
    Official site →Compare models

  • Gemini 2.0 Pro

    Google · Released 2025-02-05

    chat

    Google's multimodal flagship with a 2M-token context window.

    Context
    2M tokens
    Pricing
    Free tier · Advanced $19.99/mo · API pay-as-you-go
    Strengths
    • Massive context
    • Native Workspace integration
    • Deep Research
    Trade-offs
    • Uneven coding benchmarks
    • Slower rollout of new features
    Official site →Compare models

  • Grok 3

    xAI · Released 2025-02-17

    chat

    xAI's frontier model — real-time web access and irreverent tone.

    Context
    128K tokens
    Pricing
    Included with X Premium+ · API from $3/$15 per 1M
    Strengths
    • Real-time X data
    • DeepSearch mode
    • Fast reasoning
    Trade-offs
    • Smaller ecosystem
    • Uneven safety training
    Official site →Compare models

  • Llama 3.3

    Meta · Released 2024-12-06

    open-source

    Meta's open-weight flagship — self-hostable and free for commercial use.

    Context
    128K tokens
    Pricing
    Open weights · Free · API from $0.20/$0.60 per 1M
    Strengths
    • Truly open weights
    • Cheap to self-host
    • 70B rivals GPT-4
    Trade-offs
    • Requires ops overhead
    • Weaker on tool use
    Official site →Compare models

  • Mistral Large 2

    Mistral · Released 2024-07-24

    chat

    European frontier model with strong multilingual and coding performance.

    Context
    128K tokens
    Pricing
    $2 / $6 per 1M tokens · Free chat via Le Chat
    Strengths
    • Strong multilingual
    • EU-based (GDPR)
    • Fast API
    Trade-offs
    • Smaller ecosystem
    • Weaker on vision
    Official site →Compare models

  • DeepSeek R1

    DeepSeek · Released 2025-01-20

    reasoning

    Open reasoning model rivaling o1 at a fraction of the cost.

    Context
    128K tokens
    Pricing
    $0.55 / $2.19 per 1M tokens · Open weights
    Strengths
    • Extremely cheap
    • Open reasoning
    • Strong math
    Trade-offs
    • China-based hosting concerns
    • Newer ecosystem
    Official site →Compare models

  • Qwen 2.5

    Alibaba · Released 2024-09-19

    open-source

    Alibaba's open-weight family — strong multilingual and coding variants.

    Context
    128K tokens
    Pricing
    Open weights · Free · Hosted API pay-as-you-go
    Strengths
    • Best Chinese-language model
    • Open coder variants
    • Cheap
    Trade-offs
    • English weaker than Llama
    • Smaller Western ecosystem
    Official site →Compare models

  • Perplexity Sonar

    Perplexity · Released 2025-01-27

    reasoning

    Search-native model with real-time web grounding and citations.

    Context
    127K tokens
    Pricing
    $1 / $1 per 1M tokens · Pro $20/mo
    Strengths
    • Real-time web
    • Cited answers
    • Fast
    Trade-offs
    • Not a general-purpose chatbot
    • Weaker without web
    Official site →Compare models

  • OpenAI o3

    OpenAI · Released 2025-04-16

    reasoning

    OpenAI's advanced reasoning model — SOTA on frontier math and coding benchmarks.

    Context
    200K tokens
    Pricing
    $10 / $40 per 1M tokens · Included with Plus/Pro
    Strengths
    • Frontier reasoning
    • SOTA on ARC-AGI-2
    • Deep tool use
    Trade-offs
    • Slower than GPT-4o
    • Higher cost per token
    Official site →Compare models