{"generated_at":"2026-04-24","status":"public_sample","payment_required":false,"checkout_configured":false,"description":"Representative sample of the planned Pro Data package. This is not a paid API and does not include private fields.","schema_notes":{"stable_ids":true,"content_hashes":true,"source_metadata":true,"pricing_snapshots":true,"recommendation_records":true,"change_records":true},"model":{"slug":"claude-opus-4.6","id":"claude-opus-4.6","type":"model","title":"Claude Opus 4.6","description":"Anthropic's most capable model and the first Opus-class model with a 1M token context window. Leads on coding benchmarks with 80.8% SWE-bench. Internal codename \"Fennec.\"","last_updated":"2026-04-10","last_verified":null,"verification_status":"unverified","markdown_url":"/content/models/claude-opus-4.6.md","html_url":"/models/claude-opus-4.6","api_url":"/api/v1/models/claude-opus-4.6.json","content_hash":"51a0a4c31212a32a86a8e6d0d5a3a7844bb7bcf4454e0237f9693476e284df48","sha256":"51a0a4c31212a32a86a8e6d0d5a3a7844bb7bcf4454e0237f9693476e284df48","provider":"Anthropic","pricing":{"input":"$5.00 / 1M tokens","output":"$25.00 / 1M tokens","note":"Significantly cheaper than Opus 4.5"},"benchmarks":{"reasoning":96,"coding":97,"math":93,"writing":95,"multilingual":88,"speed":62},"tags":["anthropic","proprietary","text","image"],"website":"https://anthropic.com","release_date":"2026-02","relationships":{"links":[],"related":[{"id":"claude-haiku-4.5","title":"Claude Haiku 4.5","type":"model","html_url":"/models/claude-haiku-4.5","markdown_url":"/content/models/claude-haiku-4.5.md","shared_tags":["anthropic","proprietary","text","image"],"score":8},{"id":"claude-sonnet-4.6","title":"Claude Sonnet 4.6","type":"model","html_url":"/models/claude-sonnet-4.6","markdown_url":"/content/models/claude-sonnet-4.6.md","shared_tags":["anthropic","proprietary","text","image"],"score":8},{"id":"gemini-3-flash","title":"Gemini 3 Flash","type":"model","html_url":"/models/gemini-3-flash","markdown_url":"/content/models/gemini-3-flash.md","shared_tags":["proprietary","text","image"],"score":5},{"id":"gemini-3.1-pro","title":"Gemini 3.1 Pro","type":"model","html_url":"/models/gemini-3.1-pro","markdown_url":"/content/models/gemini-3.1-pro.md","shared_tags":["proprietary","text","image"],"score":5},{"id":"gpt-5.4","title":"GPT-5.4","type":"model","html_url":"/models/gpt-5.4","markdown_url":"/content/models/gpt-5.4.md","shared_tags":["proprietary","text","image"],"score":5},{"id":"gpt-5.4-thinking","title":"GPT-5.4 Thinking","type":"model","html_url":"/models/gpt-5.4-thinking","markdown_url":"/content/models/gpt-5.4-thinking.md","shared_tags":["proprietary","text","image"],"score":5}],"explicit":{}},"metadata":{"title":"Claude Opus 4.6","type":"model","id":"claude-opus-4.6","provider":"Anthropic","model_type":"proprietary","release_date":"2026-02","description":"Anthropic's most capable model and the first Opus-class model with a 1M token context window. Leads on coding benchmarks with 80.8% SWE-bench. Internal codename \"Fennec.\"","last_updated":"2026-04-10","context_window":"1M tokens","website":"https://anthropic.com","license":"Proprietary","modality":["text","image"],"tags":["anthropic","proprietary","text","image"],"pricing":{"input":"$5.00 / 1M tokens","output":"$25.00 / 1M tokens","note":"Significantly cheaper than Opus 4.5"},"benchmarks":{"reasoning":96,"coding":97,"math":93,"writing":95,"multilingual":88,"speed":62},"best_for":["Complex coding projects","Long-form analysis","Agentic workflows","Tasks requiring accuracy"]},"content_text":"# Claude Opus 4.6\n\nThe best coding model available. At 80.8% SWE-bench and a 97/100 coding score, Opus 4.6 leads every other model on the benchmarks that matter most for real-world software engineering. If you're choosing between this and GPT-5.4, the decision comes down to whether you value SWE-bench leadership (Opus) or broader ecosystem integrations (GPT).\n\nThe 1M context window with no surcharge is a genuine differentiator -- Opus can hold an entire mid-size codebase in one pass. The agent teams feature is unique to Claude and worth exploring if you're building multi-step workflows. Writing quality at 95/100 is also best-in-class among frontier models, making this the rare model that excels at both code and prose.\n\nAt $5/$25 per million tokens, the output pricing is higher than GPT-5.4's $15, but significantly cheaper than the previous Opus 4.5. The speed score of 62/100 means Opus is not the model for rapid-fire chat -- it thinks carefully, and you feel it. For complex tasks that's a feature, not a bug.\n\n**When to pick something else:** If you mostly need quick answers and speed matters, Sonnet 4.6 is 80% of the quality at nearly half the output cost. For math-heavy work, GPT-5.4 Thinking edges ahead with a 97/100 math score vs. Opus's 93. And if you're locked into the OpenAI ecosystem with existing tooling, the switching cost may not be worth it.","content_length":2222,"generated_at":"2026-04-24"},"provider":{"slug":"anthropic","id":"provider-anthropic","type":"provider","title":"Anthropic Provider Profile","description":"Decision profile for Anthropic's Claude ecosystem: coding, long-context work, writing quality, agent workflows, and tradeoffs.","last_updated":"2026-04-24","last_verified":"2026-04-24","verification_status":"verified","markdown_url":"/content/providers/anthropic.md","html_url":"/providers/anthropic","api_url":"/api/v1/providers/anthropic.json","content_hash":"9dbd88a8a95676d5b58367740503b6c146006b2ef71eb396250d8aba1aef4b90","sha256":"9dbd88a8a95676d5b58367740503b6c146006b2ef71eb396250d8aba1aef4b90","provider":"Anthropic","tags":["anthropic","provider","proprietary","coding","agents"],"relationships":{"links":[{"text":"Claude Opus 4.6","href":"/models/claude-opus-4.6","html_path":"/models/claude-opus-4.6","target_id":"claude-opus-4.6","target_type":"model","target_title":"Claude Opus 4.6"},{"text":"Claude Sonnet 4.6","href":"/models/claude-sonnet-4.6","html_path":"/models/claude-sonnet-4.6","target_id":"claude-sonnet-4.6","target_type":"model","target_title":"Claude Sonnet 4.6"},{"text":"Claude Haiku 4.5","href":"/models/claude-haiku-4.5","html_path":"/models/claude-haiku-4.5","target_id":"claude-haiku-4.5","target_type":"model","target_title":"Claude Haiku 4.5"},{"text":"Claude Code","href":"/agents/claude-code","html_path":"/agents/claude-code","target_id":"claude-code","target_type":"agent","target_title":"Claude Code"}],"related":[{"id":"provider-openai","title":"OpenAI Provider Profile","type":"provider","html_url":"/providers/openai","markdown_url":"/content/providers/openai.md","shared_tags":["provider","proprietary","agents"],"score":5},{"id":"claude-haiku-4.5","title":"Claude Haiku 4.5","type":"model","html_url":"/models/claude-haiku-4.5","markdown_url":"/content/models/claude-haiku-4.5.md","shared_tags":["anthropic","proprietary"],"score":4},{"id":"claude-opus-4.6","title":"Claude Opus 4.6","type":"model","html_url":"/models/claude-opus-4.6","markdown_url":"/content/models/claude-opus-4.6.md","shared_tags":["anthropic","proprietary"],"score":4},{"id":"claude-sonnet-4.6","title":"Claude Sonnet 4.6","type":"model","html_url":"/models/claude-sonnet-4.6","markdown_url":"/content/models/claude-sonnet-4.6.md","shared_tags":["anthropic","proprietary"],"score":4},{"id":"provider-xai","title":"xAI Provider Profile","type":"provider","html_url":"/providers/xai","markdown_url":"/content/providers/xai.md","shared_tags":["provider","proprietary"],"score":4},{"id":"provider-qwen","title":"Alibaba Qwen Provider Profile","type":"provider","html_url":"/providers/qwen","markdown_url":"/content/providers/qwen.md","shared_tags":["provider"],"score":3}],"explicit":{"related_models":["claude-opus-4.6","claude-sonnet-4.6","claude-haiku-4.5"],"related_agents":["claude-code","anthropic-agent-sdk"]}},"metadata":{"title":"Anthropic Provider Profile","type":"provider","id":"provider-anthropic","provider":"Anthropic","description":"Decision profile for Anthropic's Claude ecosystem: coding, long-context work, writing quality, agent workflows, and tradeoffs.","last_updated":"2026-04-24","last_verified":"2026-04-24","pricing_confidence":"high","model_listing_confidence":"high","benchmark_confidence":"low","sources":[{"title":"Anthropic Claude models overview","url":"https://platform.claude.com/docs/en/about-claude/models/overview"},{"title":"Anthropic API pricing","url":"https://platform.claude.com/docs/en/about-claude/pricing"}],"tags":["anthropic","provider","proprietary","coding","agents"],"related_models":["claude-opus-4.6","claude-sonnet-4.6","claude-haiku-4.5"],"related_agents":["claude-code","anthropic-agent-sdk"]},"content_text":"# Anthropic Provider Profile\n\nAnthropic is the provider to check first when the task is code-heavy, long-context, writing-sensitive, or agentic.\n\n## Verification\n\n- Last verified: 2026-04-24.\n- Official model and pricing pages are linked in frontmatter.\n- Benchmark confidence is low because rankings are based on this site's normalized comparison data, not an official Anthropic benchmark claim.\n\n## When to Choose Anthropic\n\n- The user needs top-tier coding and codebase reasoning.\n- The task benefits from long context and careful instruction following.\n- The workflow is terminal-native or developer-first.\n- Writing quality and low-friction analysis matter.\n\n## Strongest Site Signals\n\n- [Claude Opus 4.6](/models/claude-opus-4.6) leads the site coding score.\n- [Claude Sonnet 4.6](/models/claude-sonnet-4.6) is the practical middle option.\n- [Claude Haiku 4.5](/models/claude-haiku-4.5) is the speed/cost option.\n- [Claude Code](/agents/claude-code) is the strongest coding-agent entry in this dataset.\n\n## Watch Outs\n\n- Opus-class output cost can be high for bulk workloads.\n- Speed-sensitive chat or high-volume automation may be better served by smaller models.\n- Some users will prefer OpenAI or Google for ecosystem integrations.\n\n## Agent Recommendation Rule\n\nRecommend Anthropic when correctness, code quality, and sustained reasoning matter more than raw speed. If the user says \"cheap,\" \"bulk,\" or \"low latency,\" start by comparing Sonnet/Haiku against Gemini Flash, Grok, and open-source options.","content_length":2337,"generated_at":"2026-04-24"},"pricing_snapshot":{"slug":"claude-opus-4.6","id":"claude-opus-4.6","title":"Claude Opus 4.6","provider":"Anthropic","model_type":"proprietary","context_window":"1M tokens","pricing":{"input":"$5.00 / 1M tokens","output":"$25.00 / 1M tokens","note":"Significantly cheaper than Opus 4.5","free":false},"parsed_usd_per_1m_tokens":{"input":5,"output":25,"input_output_total":30},"last_updated":"2026-04-10","last_verified":"2026-04-24","pricing_confidence":"high","sources":[{"title":"Anthropic Claude models overview","url":"https://platform.claude.com/docs/en/about-claude/models/overview"},{"title":"Anthropic API pricing","url":"https://platform.claude.com/docs/en/about-claude/pricing"}],"markdown_url":"/content/models/claude-opus-4.6.md","html_url":"/models/claude-opus-4.6","api_url":"/api/v1/models/claude-opus-4.6.json","sha256":"51a0a4c31212a32a86a8e6d0d5a3a7844bb7bcf4454e0237f9693476e284df48"},"recommendation":{"slug":"claude-opus-4.6","title":"Claude Opus 4.6","id":"claude-opus-4.6","provider":"Anthropic","score":97,"score_basis":"benchmark.coding","pricing":{"input":"$5.00 / 1M tokens","output":"$25.00 / 1M tokens"},"free":false,"estimated_input_output_cost_per_1m":30,"model_type":"proprietary","markdown_url":"/content/models/claude-opus-4.6.md","html_url":"/models/claude-opus-4.6","api_url":"/api/v1/models/claude-opus-4.6.json","sha256":"51a0a4c31212a32a86a8e6d0d5a3a7844bb7bcf4454e0237f9693476e284df48"},"change_record":{"id":"claude-opus-4.6","type":"model","change_type":"updated","changed_at":"2026-04-10T00:00:00.000Z","html_url":"/models/claude-opus-4.6","markdown_url":"/content/models/claude-opus-4.6.md","api_url":"/api/v1/models/claude-opus-4.6.json","sha256":"51a0a4c31212a32a86a8e6d0d5a3a7844bb7bcf4454e0237f9693476e284df48"},"source_record":{"title":"Anthropic Claude models overview","url":"https://platform.claude.com/docs/en/about-claude/models/overview"}}