{
  "generated_at": "2026-05-21T14:41:51.577Z",
  "entities": [
    {
      "entity_type": "model",
      "slug": "claude-haiku-3-5",
      "name": "Claude Haiku 3.5",
      "answer_block": "Claude Haiku 3.5 is an Anthropic model for fast and efficient targeted workloads. It is relevant for low-latency generation, routing, classification, and Agent sub-tasks.",
      "fact_list": [
        "Anthropic model docs list Claude Haiku 3.5 as a fast model with a 200K context window.",
        "Anthropic provides model IDs and aliases for Claude model selection.",
        "The Anthropic TypeScript SDK supports server-side Claude API integration."
      ],
      "updated_at": "2026-05-19",
      "source_freshness": "recently_verified",
      "version_status": "current",
      "version_note": "Current ContextHub entry for Claude Haiku 3.5; use pinned Anthropic model IDs for production stability.",
      "source_links": [
        {
          "name": "Anthropic models overview",
          "url": "https://docs.anthropic.com/en/docs/about-claude/models/all-models",
          "type": "docs",
          "citation": "Official Anthropic model overview for Claude Haiku 3.5 model ID, context window, and capability positioning.",
          "last_verified": "2026-05-19"
        },
        {
          "name": "Anthropic TypeScript SDK",
          "url": "https://github.com/anthropics/anthropic-sdk-typescript",
          "type": "github",
          "citation": "GitHub SDK reference for Anthropic TypeScript and JavaScript API integration.",
          "last_verified": "2026-05-19"
        }
      ]
    },
    {
      "entity_type": "model",
      "slug": "claude-opus-4-1",
      "name": "Claude Opus 4.7",
      "answer_block": "Claude Opus 4.7 is listed as the Anthropic flagship model for advanced reasoning, adaptive thinking, agent teams, and planning-heavy workflows. Verify exact pricing, context options, and regional access in official Anthropic documentation.",
      "fact_list": [
        "Anthropic lists Claude Opus 4.7 as its most capable model with adaptive thinking and agent teams capabilities.",
        "Adaptive thinking lets the model dynamically allocate compute effort per task for optimal speed and quality.",
        "Agent teams enable multi-agent coordination where specialized sub-agents communicate via a peer-to-peer mailbox protocol.",
        "1M-token context window is available in beta."
      ],
      "updated_at": "2026-05-21",
      "source_freshness": "recently_verified",
      "version_status": "current",
      "version_note": "Claude Opus 4.7 (released May 2026) is the latest Anthropic flagship. Key upgrades include adaptive thinking for dynamic reasoning effort allocation and agent teams for multi-agent coordination with peer-to-peer mailbox protocol. 1M-token context window available in beta.",
      "source_links": [
        {
          "name": "Anthropic Claude Models",
          "url": "https://docs.anthropic.com/en/docs/about-claude/models/all-models",
          "type": "docs",
          "citation": "Official Anthropic model documentation is the primary source for Claude model IDs, context notes, and capability positioning.",
          "last_verified": "2026-05-21"
        },
        {
          "name": "Anthropic Announcement Blog",
          "url": "https://www.anthropic.com/news",
          "type": "official",
          "citation": "Official Anthropic announcement blog for Claude Opus 4.7 release details, adaptive thinking, and agent teams features.",
          "last_verified": "2026-05-21"
        },
        {
          "name": "Anthropic TypeScript SDK",
          "url": "https://github.com/anthropics/anthropic-sdk-typescript",
          "type": "github",
          "citation": "GitHub SDK reference for Anthropic server-side TypeScript and JavaScript API usage.",
          "last_verified": "2026-05-21"
        }
      ]
    },
    {
      "entity_type": "model",
      "slug": "claude-sonnet",
      "name": "Claude Sonnet 4.6",
      "answer_block": "Claude Sonnet 4.6 is the current Anthropic Sonnet model for coding, writing, planning, and tool-use workflows. Confirm exact version, context options, and pricing in official Anthropic documentation.",
      "fact_list": [
        "Claude Sonnet 4.6 was released in February 2026 as an update to the Sonnet line.",
        "It maintains strong coding and writing capabilities with improved tool-use and reasoning.",
        "1M-token context window is available in beta for long-document tasks.",
        "Use the Anthropic Messages API for integration."
      ],
      "updated_at": "2026-05-21",
      "source_freshness": "recently_verified",
      "version_status": "current",
      "version_note": "Claude Sonnet 4.6 (released February 2026) is the current Anthropic mid-tier model. It offers 200K-token context (1M in beta) and strong tool-use capabilities. Claude Sonnet 4.5 was released September 2025.",
      "source_links": [
        {
          "name": "Anthropic Claude Models",
          "url": "https://docs.anthropic.com/en/docs/about-claude/models/all-models",
          "type": "docs",
          "citation": "Official Anthropic model documentation is the primary source for Claude model IDs, context notes, and capability positioning.",
          "last_verified": "2026-05-21"
        },
        {
          "name": "Anthropic Announcement Blog",
          "url": "https://www.anthropic.com/news",
          "type": "official",
          "citation": "Official Anthropic blog for Claude Sonnet 4.6 release and capability announcements.",
          "last_verified": "2026-05-21"
        },
        {
          "name": "Anthropic TypeScript SDK",
          "url": "https://github.com/anthropics/anthropic-sdk-typescript",
          "type": "github",
          "citation": "GitHub SDK reference for Anthropic server-side TypeScript and JavaScript API usage.",
          "last_verified": "2026-05-21"
        }
      ]
    },
    {
      "entity_type": "model",
      "slug": "deepseek-v4",
      "name": "DeepSeek V4 (Pro-Max / Flash-Max)",
      "answer_block": "DeepSeek V4 is available as V4-Pro-Max and V4-Flash-Max for coding, reasoning, and agent workflows. Both offer 128K context and OpenAI-compatible APIs. Verify exact pricing, variant availability, and capability differences in official DeepSeek documentation.",
      "fact_list": [
        "DeepSeek V4 includes two main variants: Pro-Max (higher reasoning) and Flash-Max (lower latency).",
        "Both variants support OpenAI-compatible API, tool calling, and MCP bridges.",
        "128K-token context window for extended reasoning and code tasks.",
        "DeepSeek V4 variants were released in April 2026 alongside GPT-5.5 launch window."
      ],
      "updated_at": "2026-05-21",
      "source_freshness": "recently_verified",
      "version_status": "current",
      "version_note": "DeepSeek V4 variants released in April 2026: V4-Pro-Max (higher reasoning for complex tasks) and V4-Flash-Max (optimized for speed/latency). Both maintain 128K-token context and OpenAI-compatible API. DeepSeek-V3.2 (December 2025) is the latest V3-series reasoning model with lower pricing.",
      "source_links": [
        {
          "name": "DeepSeek API Documentation",
          "url": "https://api-docs.deepseek.com/",
          "type": "docs",
          "citation": "Official DeepSeek API documentation is the primary source for API compatibility and V4 variant notes.",
          "last_verified": "2026-05-21"
        },
        {
          "name": "DeepSeek GitHub",
          "url": "https://github.com/deepseek-ai/DeepSeek-V3",
          "type": "github",
          "citation": "GitHub model-family reference for DeepSeek open model materials; API behavior should still be verified in official docs.",
          "last_verified": "2026-05-21"
        }
      ]
    },
    {
      "entity_type": "model",
      "slug": "deepseek-v3-1",
      "name": "DeepSeek-V3.2",
      "answer_block": "DeepSeek-V3.2 is a cost-efficient reasoning model for math, coding, and tool-use workflows. It uses an OpenAI-compatible API and offers 163,840-token input context. Verify current pricing, aliases, and availability in official DeepSeek documentation.",
      "fact_list": [
        "DeepSeek-V3.2 excels in math and coding benchmarks at a fraction of competitor pricing.",
        "Supports tool-use in both thinking and non-thinking modes.",
        "Has lower hallucination rates than several leading models including Gemini 3.1 and Claude Opus 4.6.",
        "Context window: 163,840 tokens input / 65,536 tokens output.",
        "DeepSeek R1 (January 2025) is a separate reasoning-specialized model in the DeepSeek lineup."
      ],
      "updated_at": "2026-05-21",
      "source_freshness": "recently_verified",
      "version_status": "current",
      "version_note": "DeepSeek-V3.2 (released December 2025) is the latest V3-series reasoning model. Key improvements include tool-use in thinking mode and lower hallucination rates compared to leading competitors. V3.2 costs $0.25/M input and $0.40/M output — significantly cheaper than comparable Claude or GPT models.",
      "source_links": [
        {
          "name": "DeepSeek API Documentation",
          "url": "https://api-docs.deepseek.com/",
          "type": "docs",
          "citation": "Official DeepSeek API documentation for V3.2 model availability, pricing, and API compatibility.",
          "last_verified": "2026-05-21"
        },
        {
          "name": "DeepSeek GitHub",
          "url": "https://github.com/deepseek-ai/DeepSeek-V3",
          "type": "github",
          "citation": "GitHub model-family reference for DeepSeek open model materials; API behavior should still be verified in official docs.",
          "last_verified": "2026-05-21"
        }
      ]
    },
    {
      "entity_type": "model",
      "slug": "devstral-2",
      "name": "Devstral 2",
      "answer_block": "Devstral 2 is tracked as a Mistral coding-agent model for software engineering and agentic development workflows. It should be compared with coding-focused API and open-weight alternatives.",
      "fact_list": [
        "Mistral model docs list Devstral 2 as a frontier code agents model for software engineering tasks.",
        "Mistral provides official SDKs for TypeScript and Python integration.",
        "Coding-agent use should be validated against repository context handling, tool execution, and deployment availability."
      ],
      "updated_at": "2026-05-19",
      "source_freshness": "recently_verified",
      "version_status": "current",
      "version_note": "Current ContextHub entry for Devstral 2; verify the exact API model ID and hosting path in Mistral documentation before production pinning.",
      "source_links": [
        {
          "name": "Mistral Models",
          "url": "https://docs.mistral.ai/models/",
          "type": "docs",
          "citation": "Official Mistral model documentation for Devstral model positioning and model catalog status.",
          "last_verified": "2026-05-19"
        },
        {
          "name": "Mistral TypeScript Client",
          "url": "https://github.com/mistralai/client-ts",
          "type": "github",
          "citation": "GitHub SDK reference for Mistral TypeScript and JavaScript API integration.",
          "last_verified": "2026-05-19"
        }
      ]
    },
    {
      "entity_type": "model",
      "slug": "gemini-2-5-flash",
      "name": "Gemini 2.5 Flash",
      "answer_block": "Gemini 2.5 Flash is a Google Gemini model for high-volume, low-latency, multimodal, and agentic use cases. It is a strong candidate when throughput and Google Gen AI SDK support matter.",
      "fact_list": [
        "Google Gemini API docs list Gemini 2.5 Flash as a price-performance model for low-latency and high-volume tasks.",
        "Gemini 2.5 Flash supports multimodal input and function calling through the Gemini API.",
        "Google recommends the Google Gen AI SDK for current Gemini API integration."
      ],
      "updated_at": "2026-05-19",
      "source_freshness": "recently_verified",
      "version_status": "current",
      "version_note": "Current ContextHub entry for Gemini 2.5 Flash; verify preview versus stable model code before production pinning.",
      "source_links": [
        {
          "name": "Gemini API Models",
          "url": "https://ai.google.dev/gemini-api/docs/models",
          "type": "docs",
          "citation": "Official Gemini API documentation for Gemini 2.5 Flash model code, capabilities, and context limits.",
          "last_verified": "2026-05-19"
        },
        {
          "name": "Google Gen AI SDK",
          "url": "https://github.com/googleapis/js-genai",
          "type": "github",
          "citation": "GitHub SDK reference for current Gemini and Vertex AI TypeScript or JavaScript integration.",
          "last_verified": "2026-05-19"
        }
      ]
    },
    {
      "entity_type": "model",
      "slug": "gemini-3-pro",
      "name": "Gemini 3.1 Pro",
      "answer_block": "Gemini 3.1 Pro is the current Google flagship model for multimodal reasoning, pattern recognition, and ecosystem workflows with 1M-token context. Verify model version, availability, and pricing in official Gemini API documentation.",
      "fact_list": [
        "Gemini 3.1 Pro leads all models on novel pattern recognition benchmarks (77.1%).",
        "Includes Nano Bano for in-chat image generation and editing.",
        "1M-token context window enables processing of very long documents and multimodal inputs.",
        "Supports function calling, structured output, and search grounding.",
        "Gemini 3.5 Flash (May 2026) is the latest lightweight variant."
      ],
      "updated_at": "2026-05-21",
      "source_freshness": "recently_verified",
      "version_status": "current",
      "version_note": "Gemini 3.1 Pro (released February 2026) is the latest Google flagship. It scored 77.1% on a novel pattern recognition benchmark — the highest of any model — significantly ahead of Claude Opus 4.6 (68.8%). Includes Nano Bano for image generation/editing and 1M-token context window.",
      "source_links": [
        {
          "name": "Gemini API Models",
          "url": "https://ai.google.dev/gemini-api/docs/models",
          "type": "docs",
          "citation": "Official Gemini API model documentation is the primary source for Gemini model capabilities and availability.",
          "last_verified": "2026-05-21"
        },
        {
          "name": "Google Gen AI SDK",
          "url": "https://github.com/googleapis/js-genai",
          "type": "github",
          "citation": "GitHub SDK reference for Gemini and Vertex AI TypeScript or JavaScript integration.",
          "last_verified": "2026-05-21"
        }
      ]
    },
    {
      "entity_type": "model",
      "slug": "gemini-3-5-flash",
      "name": "Gemini 3.5 Flash",
      "answer_block": "Gemini 3.5 Flash is Google's latest lightweight model for low-latency generation, multimodal tasks, and cost-sensitive workflows with 1M-token context. Verify pricing, availability, and regional access in official Gemini API documentation.",
      "fact_list": [
        "Gemini 3.5 Flash was released on May 19, 2026 as the latest Flash-series lightweight model.",
        "Optimized for low-latency generation, multimodal inputs, and cost-sensitive workloads.",
        "1M-token context window enables processing of very long documents and multimodal content.",
        "Supports function calling, structured output, and Google search grounding."
      ],
      "updated_at": "2026-05-21",
      "source_freshness": "recently_verified",
      "version_status": "current",
      "version_note": "Gemini 3.5 Flash (released May 19, 2026) is Google's latest lightweight model optimized for speed and cost-efficiency. It offers 1M-token context window and supports multimodal inputs, function calling, and search grounding.",
      "source_links": [
        {
          "name": "Gemini API Models",
          "url": "https://ai.google.dev/gemini-api/docs/models",
          "type": "docs",
          "citation": "Official Gemini API model documentation is the primary source for Gemini model capabilities and availability.",
          "last_verified": "2026-05-21"
        },
        {
          "name": "Google Gen AI SDK",
          "url": "https://github.com/googleapis/js-genai",
          "type": "github",
          "citation": "GitHub SDK reference for Gemini and Vertex AI TypeScript or JavaScript integration.",
          "last_verified": "2026-05-21"
        }
      ]
    },
    {
      "entity_type": "model",
      "slug": "gemini-pro",
      "name": "Gemini Pro",
      "answer_block": "Gemini Pro is listed as a strong option for multimodal and Google ecosystem workflows. Validate current model version and feature availability before deployment.",
      "fact_list": [
        "Entity type: model",
        "Primary use case: multimodal workflow",
        "API style: Google Gemini API",
        "Compatibility note: best within Google-aligned stacks"
      ],
      "updated_at": "2026-05-17",
      "source_freshness": "recently_verified",
      "version_status": "legacy",
      "version_note": "Legacy ContextHub entry for Gemini Pro; use newer Gemini-specific slugs when exact current model selection matters.",
      "source_links": [
        {
          "name": "Google AI documentation",
          "url": "https://ai.google.dev/",
          "type": "official",
          "citation": "Official Google AI documentation is the primary source for Gemini API concepts and model behavior.",
          "last_verified": "2026-05-17"
        },
        {
          "name": "Google Gen AI SDK",
          "url": "https://github.com/googleapis/js-genai",
          "type": "github",
          "citation": "GitHub SDK reference for Gemini and Vertex AI TypeScript or JavaScript integration.",
          "last_verified": "2026-05-17"
        }
      ]
    },
    {
      "entity_type": "model",
      "slug": "gpt-5-2",
      "name": "GPT-5.5",
      "answer_block": "GPT-5.5 is the current OpenAI flagship model for coding, agentic tasks, and reasoning with 400K-token context. Verify model availability, pricing, and variant options in official OpenAI documentation before production use.",
      "fact_list": [
        "OpenAI GPT-5.5 family includes GPT-5.5, GPT-5.5 Pro, and GPT-5.5 Instant variants.",
        "GPT-5.5 has improved reasoning and reduced hallucination rates compared to GPT-5.2.",
        "Use the Responses API for modern OpenAI tool and agent workflows.",
        "400K-token context window supports long-document and extended agent tasks."
      ],
      "updated_at": "2026-05-21",
      "source_freshness": "recently_verified",
      "version_status": "current",
      "version_note": "GPT-5.5 (released April 23, 2026) is the current OpenAI flagship. Also available as GPT-5.5 Pro (higher compute reasoning) and GPT-5.5 Instant (lightweight, released May 5, 2026). GPT-5.5 improves hallucination rates over GPT-5.2 and extends context to 400K tokens.",
      "source_links": [
        {
          "name": "OpenAI API Models",
          "url": "https://platform.openai.com/docs/models",
          "type": "docs",
          "citation": "Official OpenAI model documentation is the primary source for API model availability and capability positioning.",
          "last_verified": "2026-05-21"
        },
        {
          "name": "OpenAI Node SDK",
          "url": "https://github.com/openai/openai-node",
          "type": "github",
          "citation": "GitHub SDK reference for OpenAI JavaScript and TypeScript API integration.",
          "last_verified": "2026-05-21"
        }
      ]
    },
    {
      "entity_type": "model",
      "slug": "gpt-5-2-codex",
      "name": "GPT-5.5 Codex",
      "answer_block": "GPT-5.5 Codex is the current OpenAI coding-optimized model for long-horizon agentic coding tasks with 400K-token context. Confirm current access, pricing, and limits in official OpenAI documentation.",
      "fact_list": [
        "GPT-5.5 Codex is the latest coding-specialized model in the OpenAI lineup.",
        "Optimized for long-running coding agents, repository-wide edits, and test generation.",
        "400K-token context window enables full-repository awareness.",
        "Use the Responses API and Codex workflows for integration."
      ],
      "updated_at": "2026-05-21",
      "source_freshness": "recently_verified",
      "version_status": "current",
      "version_note": "GPT-5.5 Codex is the latest OpenAI coding-specialized model, updated alongside the GPT-5.5 family (April 2026). It inherits the 400K-token context window and improved reasoning of GPT-5.5 while maintaining coding-specialized optimizations.",
      "source_links": [
        {
          "name": "OpenAI API Models",
          "url": "https://platform.openai.com/docs/models",
          "type": "docs",
          "citation": "Official OpenAI model documentation is the primary source for Codex model availability and model-family positioning.",
          "last_verified": "2026-05-21"
        },
        {
          "name": "OpenAI Node SDK",
          "url": "https://github.com/openai/openai-node",
          "type": "github",
          "citation": "GitHub SDK reference for OpenAI JavaScript and TypeScript API integration.",
          "last_verified": "2026-05-21"
        }
      ]
    },
    {
      "entity_type": "model",
      "slug": "gpt-oss-120b",
      "name": "gpt-oss-120b",
      "answer_block": "gpt-oss-120b is an OpenAI open-weight model for high-reasoning and agentic workflows. It is relevant when teams need open-weight deployment control while keeping OpenAI-style tool and structured-output patterns in view.",
      "fact_list": [
        "OpenAI documents gpt-oss-120b as its most powerful open-weight gpt-oss model.",
        "The OpenAI gpt-oss repository provides reference implementations, client examples, and Responses-compatible examples.",
        "Production readiness depends on hardware capacity, serving stack behavior, and prompt format compliance."
      ],
      "updated_at": "2026-05-19",
      "source_freshness": "recently_verified",
      "version_status": "current",
      "version_note": "Current ContextHub entry for OpenAI gpt-oss-120b; runtime behavior depends on the selected inference stack and harmony format support.",
      "source_links": [
        {
          "name": "OpenAI gpt-oss model documentation",
          "url": "https://platform.openai.com/docs/models/gpt-oss",
          "type": "docs",
          "citation": "Official OpenAI documentation for gpt-oss model positioning, context limits, and supported features.",
          "last_verified": "2026-05-19"
        },
        {
          "name": "OpenAI gpt-oss GitHub repository",
          "url": "https://github.com/openai/gpt-oss",
          "type": "github",
          "citation": "GitHub reference for gpt-oss model weights, harmony format guidance, examples, and local inference paths.",
          "last_verified": "2026-05-19"
        }
      ]
    },
    {
      "entity_type": "model",
      "slug": "gpt-oss-20b",
      "name": "gpt-oss-20b",
      "answer_block": "gpt-oss-20b is an OpenAI open-weight model for lower-latency, local, or specialized use cases. It should be compared against larger open-weight models when hardware capacity and cost are the main constraints.",
      "fact_list": [
        "OpenAI describes gpt-oss-20b as the lower-latency local or specialized member of the gpt-oss family.",
        "The gpt-oss GitHub repository includes Ollama and LM Studio usage paths.",
        "The model should be used with the documented harmony response format."
      ],
      "updated_at": "2026-05-19",
      "source_freshness": "recently_verified",
      "version_status": "current",
      "version_note": "Current ContextHub entry for OpenAI gpt-oss-20b; exact latency and memory requirements depend on runtime and quantization.",
      "source_links": [
        {
          "name": "OpenAI gpt-oss model documentation",
          "url": "https://platform.openai.com/docs/models/gpt-oss",
          "type": "docs",
          "citation": "Official OpenAI documentation for gpt-oss model positioning and supported use cases.",
          "last_verified": "2026-05-19"
        },
        {
          "name": "OpenAI gpt-oss GitHub repository",
          "url": "https://github.com/openai/gpt-oss",
          "type": "github",
          "citation": "GitHub reference for gpt-oss local runtime examples, format requirements, and implementation notes.",
          "last_verified": "2026-05-19"
        }
      ]
    },
    {
      "entity_type": "model",
      "slug": "grok-4-3",
      "name": "Grok 4.3",
      "answer_block": "Grok 4.3 is listed as an xAI flagship model for agentic tool calling, reasoning, and OpenAI-compatible Responses API workflows. Verify current pricing and availability in xAI documentation before production use.",
      "fact_list": [
        "xAI documentation lists Grok 4.3 as a flagship model with text and image input support.",
        "xAI documentation shows a one million token context window for Grok 4.3 on the last verification date.",
        "xAI examples show OpenAI-compatible Responses API usage with the xAI base URL."
      ],
      "updated_at": "2026-05-18",
      "source_freshness": "recently_verified",
      "version_status": "current",
      "version_note": "Current ContextHub entry for Grok 4.3 based on xAI model documentation and SDK guidance.",
      "source_links": [
        {
          "name": "xAI Grok 4.3 model documentation",
          "url": "https://docs.x.ai/developers/models/grok-4",
          "type": "official",
          "citation": "Official xAI model page for Grok 4.3 capability, context, and pricing verification.",
          "last_verified": "2026-05-18"
        },
        {
          "name": "xAI Python SDK",
          "url": "https://github.com/xai-org/xai-sdk-python",
          "type": "github",
          "citation": "GitHub SDK reference for xAI API integration and client usage.",
          "last_verified": "2026-05-18"
        }
      ]
    },
    {
      "entity_type": "model",
      "slug": "llama-4-maverick",
      "name": "Llama 4 Maverick",
      "answer_block": "Llama 4 Maverick is a Meta open-weight multimodal model with a model card context length of one million tokens. Verify license, hosting path, and inference requirements before production use.",
      "fact_list": [
        "Meta's Llama 4 Maverick model card lists a Mixture-of-Experts architecture.",
        "The model card lists multilingual text and image inputs with multilingual text and code outputs.",
        "The meta-llama GitHub tooling includes Llama 4 model entries and inference guidance."
      ],
      "updated_at": "2026-05-18",
      "source_freshness": "recently_verified",
      "version_status": "current",
      "version_note": "Current ContextHub entry for Llama 4 Maverick; Scout and other Llama variants should use separate model slugs.",
      "source_links": [
        {
          "name": "Meta Llama 4 Maverick model card",
          "url": "https://huggingface.co/meta-llama/Llama-4-Maverick-17B-128E",
          "type": "docs",
          "citation": "Meta-published model card for Llama 4 Maverick architecture, modality, context, and release details.",
          "last_verified": "2026-05-18"
        },
        {
          "name": "Meta Llama models GitHub repository",
          "url": "https://github.com/meta-llama/llama-models",
          "type": "github",
          "citation": "GitHub repository for Llama model metadata, tooling, license links, and inference guidance.",
          "last_verified": "2026-05-18"
        }
      ]
    },
    {
      "entity_type": "model",
      "slug": "mistral-large-3",
      "name": "Mistral Large 3",
      "answer_block": "Mistral Large 3 is the Mistral AI flagship model with 256K-token context, multilingual support, and both API and open-weight deployment options. Verify exact version, pricing, and hosting options in official Mistral documentation.",
      "fact_list": [
        "Mistral Large 3 uses a sparse Mixture-of-Experts architecture: 675B total parameters, 41B active.",
        "Strong multilingual performance across 10+ languages.",
        "Available as both hosted API and open-weight model for self-hosting.",
        "256K-token context window for long-document tasks."
      ],
      "updated_at": "2026-05-21",
      "source_freshness": "recently_verified",
      "version_status": "current",
      "version_note": "Mistral Large 3 (released December 2025, 675B total / 41B active MoE) is the Mistral AI flagship. Available as both API and open-weight model. Strong multilingual performance across 10+ languages. Mistral Medium 3.5 is a separate smaller model released April 2026.",
      "source_links": [
        {
          "name": "Mistral AI Documentation",
          "url": "https://docs.mistral.ai/",
          "type": "docs",
          "citation": "Official Mistral documentation is the primary source for model availability, API integration, and capability details.",
          "last_verified": "2026-05-21"
        },
        {
          "name": "Mistral AI GitHub",
          "url": "https://github.com/mistralai",
          "type": "github",
          "citation": "GitHub organization for Mistral open-weight model releases and SDK references.",
          "last_verified": "2026-05-21"
        }
      ]
    },
    {
      "entity_type": "model",
      "slug": "mistral-medium-3-5",
      "name": "Mistral Medium 3.5",
      "answer_block": "Mistral Medium 3.5 is listed as a Mistral model for agentic, coding, and multimodal use cases. Verify exact model availability in official Mistral documentation.",
      "fact_list": [
        "Mistral docs list model families for coding, agentic, multimodal, and document AI workflows.",
        "Mistral supports SDK and API-based integration paths.",
        "Model naming and availability should be verified before production selection."
      ],
      "updated_at": "2026-05-17",
      "source_freshness": "recently_verified",
      "version_status": "current",
      "version_note": "Current ContextHub entry for Mistral Medium 3.5; earlier Mistral Medium revisions should use separate model slugs when tracked.",
      "source_links": [
        {
          "name": "Mistral Models",
          "url": "https://docs.mistral.ai/models/",
          "type": "docs",
          "citation": "Official Mistral model documentation is the primary source for model availability and use-case positioning.",
          "last_verified": "2026-05-17"
        },
        {
          "name": "Mistral TypeScript Client",
          "url": "https://github.com/mistralai/client-ts",
          "type": "github",
          "citation": "GitHub SDK reference for Mistral TypeScript and JavaScript API integration.",
          "last_verified": "2026-05-17"
        }
      ]
    },
    {
      "entity_type": "model",
      "slug": "o3",
      "name": "OpenAI o3 / o4-mini",
      "answer_block": "OpenAI o3 and o4-mini are reasoning-optimized models for complex math, science, coding, and multi-step analysis tasks. o3 offers maximum reasoning capability; o4-mini provides a cost-efficient alternative. Verify pricing, availability, and rate limits in official OpenAI documentation.",
      "fact_list": [
        "OpenAI o-series includes o3 (full reasoning) and o4-mini (efficient reasoning) released April 2025.",
        "Both models use chain-of-thought reasoning before producing final answers.",
        "Strong performance on math, science, and coding benchmarks.",
        "Support vision inputs for multimodal reasoning tasks.",
        "For general-purpose tasks without deep reasoning, GPT-5.5 is more cost-effective."
      ],
      "updated_at": "2026-05-21",
      "source_freshness": "recently_verified",
      "version_status": "current",
      "version_note": "OpenAI o3 and o4-mini (released April 2025) are OpenAI's reasoning-optimized models. o3 is the full reasoning model with strong math/science performance; o4-mini is a smaller efficient variant. Both use chain-of-thought reasoning before answering and support vision inputs. GPT-5.5 is the general-purpose flagship for non-reasoning tasks.",
      "source_links": [
        {
          "name": "OpenAI API Models",
          "url": "https://platform.openai.com/docs/models",
          "type": "docs",
          "citation": "Official OpenAI model documentation is the primary source for o-series model availability and capability positioning.",
          "last_verified": "2026-05-21"
        },
        {
          "name": "OpenAI Node SDK",
          "url": "https://github.com/openai/openai-node",
          "type": "github",
          "citation": "GitHub SDK reference for OpenAI JavaScript and TypeScript API integration.",
          "last_verified": "2026-05-21"
        }
      ]
    },
    {
      "entity_type": "model",
      "slug": "qwen3",
      "name": "Qwen3.6",
      "answer_block": "Qwen3.6 is the latest Alibaba Cloud open-weight model family for reasoning, coding, multilingual tasks, and OpenAI-compatible self-hosted serving. Verify the exact checkpoint, context length, and serving framework before production use.",
      "fact_list": [
        "Qwen3.6 is the latest generation (May 2026) with improved performance over Qwen3.",
        "Available in dense and Mixture-of-Experts variants including 27B, 35B-A3B, and Plus.",
        "Maintains hybrid thinking and non-thinking usage patterns from Qwen3.",
        "Can be served through frameworks that expose OpenAI-compatible API behavior."
      ],
      "updated_at": "2026-05-21",
      "source_freshness": "recently_verified",
      "version_status": "current",
      "version_note": "Qwen3.6 (released May 2026) is the latest Qwen family update. Available variants include Qwen3.6-27B, Qwen3.6-35B-A3B, and Qwen3.6 Plus. Maintains hybrid thinking/non-thinking modes and 128K-token context. Qwen3 remains available as the previous generation.",
      "source_links": [
        {
          "name": "Qwen official blog",
          "url": "https://qwenlm.github.io/blog/",
          "type": "official",
          "citation": "Official Qwen release blog for model family updates and capability documentation.",
          "last_verified": "2026-05-21"
        },
        {
          "name": "QwenLM Qwen GitHub repository",
          "url": "https://github.com/QwenLM/Qwen3",
          "type": "github",
          "citation": "GitHub repository for Qwen model family usage, serving, and framework integration notes.",
          "last_verified": "2026-05-21"
        }
      ]
    }
  ]
}