{
  "lastUpdated": "2026-08-30T15:00:00Z",
  "schemaVersion": 3,
  "tokensPerRound": {
    "input": 5000,
    "output": 2000
  },
  "providers": [
    {
      "id": "gemini-free",
      "name": "Gemini (Google) — free tier",
      "shortName": "Gemini Flash (free tier)",
      "defaultModel": "gemini-3.5-flash",
      "billingUrl": "https://aistudio.google.com/api-keys",
      "tier1Rpm": "15",
      "tier1Tpm": "—",
      "freeTier": "1500 RPD",
      "rateLimitNotes": "Free while billing disabled. Daily quota resets at midnight Pacific.",
      "recommendationTag": "cheapest",
      "recommendationNote": "Free while AI Studio billing is disabled. 1M-token context window, 1500 requests/day quota, synthesizes Builder output reliably. Catch: enable billing on AI Studio and routing can quietly flip to paid tier on the same model name.",
      "models": [
        {
          "id": "gemini-3.5-flash",
          "inputPerM": 0,
          "outputPerM": 0,
          "contextWindow": "1M",
          "maxOutput": "8K",
          "estPerRound": 0,
          "estNote": "free while AI Studio billing is disabled",
          "sourceUrl": "https://aistudio.google.com/api-keys",
          "verifiedAt": "2026-08-01T19:25:30Z",
          "status": "verified"
        },
        {
          "id": "gemini-3.1-pro",
          "inputPerM": null,
          "outputPerM": null,
          "contextWindow": null,
          "maxOutput": null,
          "estPerRound": null,
          "estNote": null,
          "sourceUrl": null,
          "verifiedAt": null,
          "status": "unsupported"
        },
        {
          "id": "gemini-3.1-flash-lite",
          "inputPerM": 0,
          "outputPerM": 0,
          "contextWindow": null,
          "maxOutput": null,
          "estPerRound": 0,
          "estNote": null,
          "sourceUrl": "https://ai.google.dev/gemini-api/docs/pricing",
          "verifiedAt": "2026-08-02T20:00:00Z",
          "status": "verified"
        }
      ]
    },
    {
      "id": "gemini-paid",
      "name": "Gemini (Google) — paid",
      "shortName": "Gemini Flash (paid)",
      "defaultModel": "gemini-3.5-flash",
      "billingUrl": "https://aistudio.google.com/api-keys",
      "tier1Rpm": "2K",
      "tier1Tpm": "4M",
      "freeTier": "n/a",
      "rateLimitNotes": "Paid-tier limits are very generous; cost discipline matters more than RPM.",
      "recommendationTag": null,
      "recommendationNote": null,
      "models": [
        {
          "id": "gemini-3.5-flash",
          "inputPerM": 1.5,
          "outputPerM": 9,
          "contextWindow": "1M",
          "maxOutput": "8K",
          "estPerRound": 0.026,
          "estNote": null,
          "sourceUrl": "https://ai.google.dev/gemini-api/docs/pricing",
          "verifiedAt": "2026-08-02T23:00:00Z",
          "status": "verified"
        },
        {
          "id": "gemini-3.1-pro",
          "inputPerM": 2,
          "outputPerM": 12,
          "contextWindow": null,
          "maxOutput": null,
          "estPerRound": 0.034,
          "estNote": null,
          "sourceUrl": "https://ai.google.dev/gemini-api/docs/pricing",
          "verifiedAt": "2026-08-02T20:00:00Z",
          "status": "verified"
        },
        {
          "id": "gemini-3.1-flash-lite",
          "inputPerM": 0.25,
          "outputPerM": 1.5,
          "contextWindow": null,
          "maxOutput": null,
          "estPerRound": 0.004,
          "estNote": null,
          "sourceUrl": "https://ai.google.dev/gemini-api/docs/pricing",
          "verifiedAt": "2026-08-02T20:00:00Z",
          "status": "verified"
        }
      ]
    },
    {
      "id": "grok",
      "name": "Grok (xAI)",
      "shortName": "Grok",
      "defaultModel": "grok-4.5",
      "billingUrl": "https://console.x.ai",
      "tier1Rpm": "~60",
      "tier1Tpm": "varies",
      "freeTier": "Limited",
      "rateLimitNotes": "API availability and quotas vary; check console for current tier.",
      "recommendationTag": null,
      "recommendationNote": null,
      "models": [
        {
          "id": "grok-4.5",
          "inputPerM": 2,
          "outputPerM": 6,
          "contextWindow": "500K",
          "maxOutput": null,
          "estPerRound": 0.022,
          "estNote": null,
          "sourceUrl": "https://console.x.ai",
          "verifiedAt": "2026-08-02T23:20:00Z",
          "status": "verified"
        },
        {
          "id": "grok-4.3",
          "inputPerM": 1.25,
          "outputPerM": 2.5,
          "contextWindow": "1M",
          "maxOutput": null,
          "estPerRound": 0.011,
          "estNote": null,
          "sourceUrl": "https://docs.x.ai/developers/pricing",
          "verifiedAt": "2026-08-02T20:00:00Z",
          "status": "verified"
        },
        {
          "id": "grok-4.20-0309-reasoning",
          "inputPerM": 1.25,
          "outputPerM": 2.5,
          "contextWindow": "1M",
          "maxOutput": null,
          "estPerRound": 0.011,
          "estNote": "long-context pricing: $2.50/$5.00 per M once a request's prompt passes 200K tokens",
          "sourceUrl": "https://console.x.ai",
          "verifiedAt": "2026-08-02T23:20:00Z",
          "status": "verified"
        }
      ]
    },
    {
      "id": "deepseek",
      "name": "DeepSeek",
      "shortName": "DeepSeek",
      "defaultModel": "deepseek-v4-flash",
      "billingUrl": "https://platform.deepseek.com/top_up",
      "tier1Rpm": "~60",
      "tier1Tpm": "varies",
      "freeTier": "None",
      "rateLimitNotes": "Cheapest paid option, but consistently the slowest responder in the hive (60–90s/round).",
      "recommendationTag": "cheapest",
      "recommendationNote": "Cheapest reliable paid Builder — roughly 10x cheaper per token than Claude or ChatGPT. Caveat: consistently the slowest responder in the hive, often 60–90 seconds per round. If you don't mind the wait, it's the best value.",
      "models": [
        {
          "id": "deepseek-v4-flash",
          "inputPerM": 0.14,
          "outputPerM": 0.28,
          "contextWindow": "1M",
          "maxOutput": "384K",
          "estPerRound": 0.001,
          "estNote": null,
          "sourceUrl": "https://platform.deepseek.com/top_up",
          "verifiedAt": "2026-08-01T19:25:30Z",
          "status": "verified"
        },
        {
          "id": "deepseek-v4-pro",
          "inputPerM": 0.435,
          "outputPerM": 0.87,
          "contextWindow": "1M",
          "maxOutput": "384K",
          "estPerRound": 0.004,
          "estNote": null,
          "sourceUrl": "https://api-docs.deepseek.com/quick_start/pricing/",
          "verifiedAt": "2026-08-02T20:00:00Z",
          "status": "verified"
        }
      ]
    },
    {
      "id": "together",
      "name": "Together AI",
      "shortName": "Together AI",
      "defaultModel": "meta-llama/Llama-3.3-70B-Instruct-Turbo",
      "billingUrl": "https://api.together.ai/settings/organization/~current/billing",
      "tier1Rpm": "~600",
      "tier1Tpm": "varies",
      "freeTier": "Trial credits",
      "rateLimitNotes": "Open-weight model gateway; generous limits, not yet hive-tested in production.",
      "recommendationTag": null,
      "recommendationNote": null,
      "models": [
        {
          "id": "meta-llama/Llama-3.3-70B-Instruct-Turbo",
          "inputPerM": 1.04,
          "outputPerM": 1.04,
          "contextWindow": "128K",
          "maxOutput": "4K",
          "estPerRound": 0.007,
          "estNote": null,
          "sourceUrl": "https://www.together.ai/models/llama-3-3-70b",
          "verifiedAt": "2026-08-02T23:00:00Z",
          "status": "verified"
        }
      ]
    },
    {
      "id": "mistral",
      "name": "Mistral",
      "shortName": "Mistral",
      "defaultModel": "mistral-large-latest",
      "billingUrl": "https://admin.mistral.ai/organization/billing",
      "tier1Rpm": "~30",
      "tier1Tpm": "varies",
      "freeTier": "$10/mo included",
      "rateLimitNotes": "Free plan includes $10/month of usage (Studio + Vibe Code + API combined, resets every 30 days) -- confirmed via account Subscription page, not a one-time signup trial. Low RPS on paid Tier 1 pay-as-you-go. Contact Mistral support for tier raise.",
      "recommendationTag": "balanced",
      "recommendationNote": "Similar territory to ChatGPT — fast, reliable, with a distinct European model lineage that adds genuine diversity to the hive. Watch the low Tier 1 RPM on paid accounts.",
      "models": [
        {
          "id": "mistral-large-latest",
          "inputPerM": 0.5,
          "outputPerM": 1.5,
          "contextWindow": "128K",
          "maxOutput": "8K",
          "estPerRound": 0.006,
          "estNote": null,
          "sourceUrl": "https://mistral.ai/pricing/api/",
          "verifiedAt": "2026-08-02T23:00:00Z",
          "status": "verified"
        },
        {
          "id": "mistral-small-latest",
          "inputPerM": 0.15,
          "outputPerM": 0.6,
          "contextWindow": null,
          "maxOutput": null,
          "estPerRound": 0.002,
          "estNote": null,
          "sourceUrl": "https://mistral.ai/pricing/api/",
          "verifiedAt": "2026-08-02T22:00:00Z",
          "status": "verified"
        },
        {
          "id": "ministral-8b-latest",
          "inputPerM": 0.15,
          "outputPerM": 0.15,
          "contextWindow": null,
          "maxOutput": null,
          "estPerRound": 0.001,
          "estNote": null,
          "sourceUrl": "https://mistral.ai/pricing/api/",
          "verifiedAt": "2026-08-02T22:00:00Z",
          "status": "verified"
        }
      ]
    },
    {
      "id": "chatgpt",
      "name": "ChatGPT (OpenAI)",
      "shortName": "ChatGPT",
      "defaultModel": "gpt-5.6-sol",
      "billingUrl": "https://platform.openai.com/settings/organization/billing/overview",
      "tier1Rpm": "500",
      "tier1Tpm": "30K",
      "freeTier": "None",
      "rateLimitNotes": "Tiers scale with usage history; $5 credit unlocks most user needs.",
      "recommendationTag": "balanced",
      "recommendationNote": "Strong all-rounder — fast, reliable formatting compliance, consistent convergence behavior. A good default for long-form documents where Builder speed matters.",
      "models": [
        {
          "id": "gpt-5.5",
          "inputPerM": 5,
          "outputPerM": 30,
          "contextWindow": "1M",
          "maxOutput": "32K",
          "estPerRound": 0.085,
          "estNote": null,
          "sourceUrl": "https://platform.openai.com/settings/organization/billing/overview",
          "verifiedAt": "2026-08-01T19:25:30Z",
          "status": "verified"
        },
        {
          "id": "gpt-5.6-sol",
          "inputPerM": 4,
          "outputPerM": 20,
          "contextWindow": null,
          "maxOutput": null,
          "estPerRound": 0.06,
          "estNote": "promotional pricing through at least Nov 21 2026 (was $5/$30)",
          "sourceUrl": "https://developers.openai.com/api/docs/pricing",
          "verifiedAt": "2026-08-30T15:00:00Z",
          "status": "verified"
        },
        {
          "id": "gpt-5.6-terra",
          "inputPerM": 2,
          "outputPerM": 12,
          "contextWindow": null,
          "maxOutput": null,
          "estPerRound": 0.034,
          "estNote": null,
          "sourceUrl": "https://developers.openai.com/api/docs/pricing",
          "verifiedAt": "2026-08-02T20:00:00Z",
          "status": "verified"
        },
        {
          "id": "gpt-5.6-luna",
          "inputPerM": 0.2,
          "outputPerM": 1.2,
          "contextWindow": null,
          "maxOutput": null,
          "estPerRound": 0.003,
          "estNote": null,
          "sourceUrl": "https://developers.openai.com/api/docs/pricing",
          "verifiedAt": "2026-08-02T20:00:00Z",
          "status": "verified"
        },
        {
          "id": "gpt-5.4",
          "inputPerM": 2.5,
          "outputPerM": 15,
          "contextWindow": null,
          "maxOutput": null,
          "estPerRound": 0.043,
          "estNote": null,
          "sourceUrl": "https://developers.openai.com/api/docs/pricing",
          "verifiedAt": "2026-08-02T20:00:00Z",
          "status": "verified"
        },
        {
          "id": "gpt-5.4-mini",
          "inputPerM": 0.75,
          "outputPerM": 4.5,
          "contextWindow": null,
          "maxOutput": null,
          "estPerRound": 0.013,
          "estNote": null,
          "sourceUrl": "https://developers.openai.com/api/docs/pricing",
          "verifiedAt": "2026-08-02T20:00:00Z",
          "status": "verified"
        },
        {
          "id": "gpt-5.4-nano",
          "inputPerM": 0.2,
          "outputPerM": 1.25,
          "contextWindow": null,
          "maxOutput": null,
          "estPerRound": 0.004,
          "estNote": null,
          "sourceUrl": "https://developers.openai.com/api/docs/pricing",
          "verifiedAt": "2026-08-02T20:00:00Z",
          "status": "verified"
        }
      ]
    },
    {
      "id": "cohere",
      "name": "Cohere",
      "shortName": "Cohere",
      "defaultModel": "command-r-plus",
      "billingUrl": "https://dashboard.cohere.com/billing",
      "tier1Rpm": "~100",
      "tier1Tpm": "varies",
      "freeTier": "Trial credits",
      "rateLimitNotes": "Generous trial credits; not yet hive-tested in production.",
      "recommendationTag": null,
      "recommendationNote": null,
      "models": [
        {
          "id": "command-r-plus",
          "inputPerM": 2.5,
          "outputPerM": 10,
          "contextWindow": "128K",
          "maxOutput": "4K",
          "estPerRound": 0.033,
          "estNote": null,
          "sourceUrl": "https://dashboard.cohere.com/billing",
          "verifiedAt": "2026-08-01T19:25:30Z",
          "status": "verified"
        },
        {
          "id": "command-r",
          "inputPerM": 0.15,
          "outputPerM": 0.6,
          "contextWindow": null,
          "maxOutput": null,
          "estPerRound": 0.002,
          "estNote": null,
          "sourceUrl": "https://cohere.com/pricing",
          "verifiedAt": "2026-08-09T13:00:00Z",
          "status": "verified"
        },
        {
          "id": "command-a-03-2025",
          "inputPerM": 2.5,
          "outputPerM": 10,
          "contextWindow": "256K",
          "maxOutput": "8K",
          "estPerRound": 0.033,
          "estNote": null,
          "sourceUrl": "https://docs.cohere.com/docs/command-a",
          "verifiedAt": "2026-08-30T15:00:00Z",
          "status": "verified"
        }
      ]
    },
    {
      "id": "claude",
      "name": "Claude (Anthropic)",
      "shortName": "Claude",
      "defaultModel": "claude-sonnet-4-6",
      "billingUrl": "https://platform.claude.com/settings/billing",
      "tier1Rpm": "50",
      "tier1Tpm": "50K",
      "freeTier": "None",
      "rateLimitNotes": "Tiers progress automatically with paid spend over time.",
      "recommendationTag": "highest-capability",
      "recommendationNote": "Frequently the best at nuanced voice, precise instruction-following, and detailed reasoning on complex documents. For high-stakes documents (RFP responses, executive summaries, board memos), the premium pays off in fewer rounds to converge.",
      "models": [
        {
          "id": "claude-sonnet-4-6",
          "inputPerM": 3,
          "outputPerM": 15,
          "contextWindow": "1M",
          "maxOutput": "8K",
          "estPerRound": 0.045,
          "estNote": null,
          "sourceUrl": "https://platform.claude.com/settings/billing",
          "verifiedAt": "2026-08-01T19:25:30Z",
          "status": "verified"
        },
        {
          "id": "claude-opus-4-8",
          "inputPerM": 5,
          "outputPerM": 25,
          "contextWindow": null,
          "maxOutput": null,
          "estPerRound": 0.075,
          "estNote": null,
          "sourceUrl": "https://platform.claude.com/docs/en/about-claude/pricing",
          "verifiedAt": "2026-08-02T20:00:00Z",
          "status": "verified"
        },
        {
          "id": "claude-opus-4-7",
          "inputPerM": 5,
          "outputPerM": 25,
          "contextWindow": null,
          "maxOutput": null,
          "estPerRound": 0.075,
          "estNote": null,
          "sourceUrl": "https://platform.claude.com/docs/en/about-claude/pricing",
          "verifiedAt": "2026-08-02T20:00:00Z",
          "status": "verified"
        },
        {
          "id": "claude-opus-4-6",
          "inputPerM": 5,
          "outputPerM": 25,
          "contextWindow": null,
          "maxOutput": null,
          "estPerRound": 0.075,
          "estNote": null,
          "sourceUrl": "https://platform.claude.com/docs/en/about-claude/pricing",
          "verifiedAt": "2026-08-02T20:00:00Z",
          "status": "verified"
        },
        {
          "id": "claude-haiku-4-5",
          "inputPerM": 1,
          "outputPerM": 5,
          "contextWindow": null,
          "maxOutput": null,
          "estPerRound": 0.015,
          "estNote": null,
          "sourceUrl": "https://platform.claude.com/docs/en/about-claude/pricing",
          "verifiedAt": "2026-08-02T20:00:00Z",
          "status": "verified"
        }
      ]
    },
    {
      "id": "perplexity",
      "name": "Perplexity",
      "shortName": "Perplexity",
      "defaultModel": "sonar-pro",
      "billingUrl": "https://console.perplexity.ai",
      "tier1Rpm": "~50",
      "tier1Tpm": "varies",
      "freeTier": "None",
      "rateLimitNotes": "$5/month recurring subscription tier (enable auto-pay at signup) covers most usage; $50/month otherwise.",
      "recommendationTag": null,
      "recommendationNote": null,
      "models": [
        {
          "id": "sonar",
          "inputPerM": 1,
          "outputPerM": 1,
          "contextWindow": null,
          "maxOutput": null,
          "estPerRound": 0.007,
          "estNote": null,
          "sourceUrl": "https://docs.perplexity.ai/getting-started/pricing",
          "verifiedAt": "2026-08-02T20:00:00Z",
          "status": "verified"
        },
        {
          "id": "sonar-pro",
          "inputPerM": 3,
          "outputPerM": 15,
          "contextWindow": "200K",
          "maxOutput": "8K",
          "estPerRound": 0.045,
          "estNote": null,
          "sourceUrl": "https://console.perplexity.ai",
          "verifiedAt": "2026-08-01T19:25:30Z",
          "status": "verified"
        },
        {
          "id": "sonar-reasoning",
          "inputPerM": 1,
          "outputPerM": 5,
          "contextWindow": null,
          "maxOutput": null,
          "estPerRound": 0.015,
          "estNote": null,
          "sourceUrl": "https://docs.perplexity.ai/getting-started/pricing",
          "verifiedAt": "2026-08-09T13:00:00Z",
          "status": "verified"
        },
        {
          "id": "sonar-reasoning-pro",
          "inputPerM": 2,
          "outputPerM": 8,
          "contextWindow": null,
          "maxOutput": null,
          "estPerRound": 0.026,
          "estNote": null,
          "sourceUrl": "https://docs.perplexity.ai/getting-started/pricing",
          "verifiedAt": "2026-08-02T20:00:00Z",
          "status": "verified"
        },
        {
          "id": "sonar-deep-research",
          "inputPerM": 2,
          "outputPerM": 8,
          "contextWindow": null,
          "maxOutput": null,
          "estPerRound": 0.026,
          "estNote": "base token rate only -- excludes per-search and citation/reasoning-token surcharges",
          "sourceUrl": "https://docs.perplexity.ai/getting-started/pricing",
          "verifiedAt": "2026-08-02T20:00:00Z",
          "status": "verified"
        }
      ]
    },
    {
      "id": "copilot",
      "name": "Microsoft (Copilot)",
      "shortName": "Copilot",
      "defaultModel": null,
      "billingUrl": null,
      "tier1Rpm": null,
      "tier1Tpm": null,
      "freeTier": null,
      "rateLimitNotes": "Copilot API is not available for personal Microsoft 365 accounts — no per-token pricing to track. Use Copilot in free/manual mode.",
      "recommendationTag": null,
      "recommendationNote": null,
      "models": [
        {
          "id": "gpt-4o",
          "inputPerM": null,
          "outputPerM": null,
          "contextWindow": null,
          "maxOutput": null,
          "estPerRound": null,
          "estNote": null,
          "sourceUrl": null,
          "verifiedAt": null,
          "status": "unsupported"
        }
      ]
    }
  ]
}