[
    {
        "name": "Claude Opus 5.5",
        "category": "proprietary",
        "creator": "Anthropic",
        "description": "The current state-of-the-art for complex agentic workflows, Opus 5.5 excels at multi-step reasoning and long-horizon planning. Its advanced self-correction capabilities make it highly reliable for critical autonomous tasks.",
        "contextWindow": "2,000,000 tokens",
        "contextValue": 2000000,
        "primaryStrength": "Agentic Coding & Intelligence",
        "sweBench": 72.0,
        "gpqa": 83.0,
        "humanEval": 92.5,
        "gaia": 74.0,
        "webarena": 62.0,
        "launch_date": "October 1, 2026",
        "costPerMillion": 12.0,
        "costLabel": "$12.00",
        "color": "from-purple-500 to-indigo-500",
        "last_updated": "2026-11-04T19:30:58.863783+00:00"
    },
    {
        "name": "GPT-6 Astra",
        "category": "proprietary",
        "creator": "OpenAI",
        "description": "OpenAI's flagship model, Astra is a generalist powerhouse with exceptional performance across all benchmarks. It is particularly strong in complex problem-solving and creative generation for agentic applications.",
        "contextWindow": "1,500,000 tokens",
        "contextValue": 1500000,
        "primaryStrength": "General Agentic Intelligence",
        "sweBench": 68.0,
        "gpqa": 80.5,
        "humanEval": 89.0,
        "gaia": 70.0,
        "webarena": 58.0,
        "launch_date": "September 15, 2026",
        "costPerMillion": 10.0,
        "costLabel": "$10.00",
        "color": "from-blue-500 to-cyan-500",
        "last_updated": "2026-11-04T19:30:58.863783+00:00"
    },
    {
        "name": "Claude Fable 5.1",
        "category": "proprietary",
        "creator": "Anthropic",
        "description": "A highly capable model optimized for nuanced understanding and robust conversational agent interactions. Fable 5.1 offers a strong balance of performance and cost-efficiency for advanced agent deployments.",
        "contextWindow": "1,000,000 tokens",
        "contextValue": 1000000,
        "primaryStrength": "Nuanced Reasoning & Conversation",
        "sweBench": 65.0,
        "gpqa": 78.0,
        "humanEval": 87.0,
        "gaia": 66.0,
        "webarena": 54.0,
        "launch_date": "August 20, 2026",
        "costPerMillion": 8.0,
        "costLabel": "$8.00",
        "color": "from-purple-400 to-pink-500",
        "last_updated": "2026-11-04T19:30:58.863783+00:00"
    },
    {
        "name": "GPT-6 Sol",
        "category": "proprietary",
        "creator": "OpenAI",
        "description": "Designed for high-throughput agentic tasks, Sol provides a strong performance-to-cost ratio. It excels in data analysis and structured output generation, making it ideal for automated data pipelines.",
        "contextWindow": "1,000,000 tokens",
        "contextValue": 1000000,
        "primaryStrength": "High-Throughput Agentic Tasks",
        "sweBench": 63.0,
        "gpqa": 76.0,
        "humanEval": 85.0,
        "gaia": 65.0,
        "webarena": 52.0,
        "launch_date": "July 1, 2026",
        "costPerMillion": 6.0,
        "costLabel": "$6.00",
        "color": "from-blue-400 to-sky-500",
        "last_updated": "2026-11-04T19:30:58.863783+00:00"
    },
    {
        "name": "Gemini 3.8 Flash",
        "category": "proprietary",
        "creator": "Google",
        "description": "Google's latest iteration of its Flash series, optimized for speed and efficiency in real-time agentic applications. It offers robust performance for interactive agents and rapid decision-making systems.",
        "contextWindow": "750,000 tokens",
        "contextValue": 750000,
        "primaryStrength": "Speed & Real-time Agentics",
        "sweBench": 60.0,
        "gpqa": 74.0,
        "humanEval": 82.0,
        "gaia": 62.0,
        "webarena": 49.0,
        "launch_date": "June 10, 2026",
        "costPerMillion": 5.0,
        "costLabel": "$5.00",
        "color": "from-red-500 to-orange-500",
        "last_updated": "2026-11-04T19:30:58.863783+00:00"
    },
    {
        "name": "Llama 4 Maverick",
        "category": "open",
        "creator": "Meta",
        "description": "Meta's leading open-source model, Maverick provides strong capabilities for custom agent development and fine-tuning. Its open nature allows for unparalleled flexibility and community-driven innovation.",
        "contextWindow": "500,000 tokens",
        "contextValue": 500000,
        "primaryStrength": "Open-source Agent Development",
        "sweBench": 55.0,
        "gpqa": 69.0,
        "humanEval": 78.0,
        "gaia": 56.0,
        "webarena": 46.0,
        "launch_date": "May 25, 2026",
        "costPerMillion": 0.0,
        "costLabel": "Self-hosted",
        "color": "from-pink-500 to-rose-500",
        "last_updated": "2026-11-04T19:30:58.863783+00:00"
    },
    {
        "name": "Qwen3.8 235B",
        "category": "proprietary",
        "creator": "Alibaba",
        "description": "Alibaba's latest large-scale model, Qwen3.8 235B demonstrates strong performance in complex reasoning and multilingual agentic tasks. It is a robust choice for enterprise-level automation.",
        "contextWindow": "600,000 tokens",
        "contextValue": 600000,
        "primaryStrength": "Enterprise Automation & Multilingual",
        "sweBench": 52.0,
        "gpqa": 67.0,
        "humanEval": 75.0,
        "gaia": 53.0,
        "webarena": 44.0,
        "launch_date": "April 1, 2026",
        "costPerMillion": 4.5,
        "costLabel": "$4.50",
        "color": "from-orange-500 to-yellow-500",
        "last_updated": "2026-11-04T19:30:58.863783+00:00"
    },
    {
        "name": "GPT-6 Luna",
        "category": "proprietary",
        "creator": "OpenAI",
        "description": "A more compact and cost-effective model in the GPT-6 series, Luna is suitable for a wide range of agentic applications. It offers solid performance for tasks requiring moderate reasoning and planning.",
        "contextWindow": "500,000 tokens",
        "contextValue": 500000,
        "primaryStrength": "Cost-effective General Agentics",
        "sweBench": 48.0,
        "gpqa": 63.0,
        "humanEval": 70.0,
        "gaia": 50.0,
        "webarena": 41.0,
        "launch_date": "March 1, 2026",
        "costPerMillion": 3.0,
        "costLabel": "$3.00",
        "color": "from-indigo-400 to-purple-400",
        "last_updated": "2026-11-04T19:30:58.863783+00:00"
    },
    {
        "name": "Llama 4 Scout",
        "category": "open",
        "creator": "Meta",
        "description": "Scout is Meta's long-context open-source model, designed for applications requiring extensive memory and information retrieval. It enables agents to operate effectively over vast amounts of data.",
        "contextWindow": "10,000,000 tokens",
        "contextValue": 10000000,
        "primaryStrength": "Long-Context Reasoning & Retrieval",
        "sweBench": 45.0,
        "gpqa": 60.0,
        "humanEval": 68.0,
        "gaia": 47.0,
        "webarena": 39.0,
        "launch_date": "February 15, 2026",
        "costPerMillion": 0.0,
        "costLabel": "Self-hosted",
        "color": "from-fuchsia-500 to-pink-400",
        "last_updated": "2026-11-04T19:30:58.863783+00:00"
    },
    {
        "name": "Meta Muse Glimmer",
        "category": "proprietary",
        "creator": "Meta",
        "description": "Glimmer is Meta's specialized model for creative and multimodal agentic tasks, focusing on generating diverse outputs. It supports agents in content creation and interactive experiences.",
        "contextWindow": "400,000 tokens",
        "contextValue": 400000,
        "primaryStrength": "Creative & Multimodal Agentics",
        "sweBench": 44.0,
        "gpqa": 58.0,
        "humanEval": 65.0,
        "gaia": 46.0,
        "webarena": 38.0,
        "launch_date": "January 10, 2026",
        "costPerMillion": 2.5,
        "costLabel": "$2.50",
        "color": "from-emerald-500 to-teal-500",
        "last_updated": "2026-11-04T19:30:58.863783+00:00"
    }
]