{
  "meta": {
    "currency": "USD",
    "generated": "2026-09-25",
    "unit": "per 1M tokens",
    "verification_policy": "verified = confirmed against a provider pricing doc or a reputable pricing index checked in September 2026. unverified = conflicting or stale sources; do not treat the listed figure as authoritative. List prices only — your bill also depends on caching, batch discounts, peak windows, thinking tokens, and tokenizer differences."
  },
  "models": [
    {
      "id": "gpt-6-astra",
      "provider": "OpenAI",
      "model": "GPT-6 Astra",
      "input_per_1m": 10.0,
      "output_per_1m": 50.0,
      "last_verified": "2026-09-25",
      "source": "https://developers.openai.com/api/docs/pricing",
      "unverified": false,
      "notes": "Flagship tier. Batch/Flex: $5/$25. Long context: $20/$75."
    },
    {
      "id": "gpt-6-sol",
      "provider": "OpenAI",
      "model": "GPT-6 Sol",
      "input_per_1m": 2.0,
      "output_per_1m": 10.0,
      "last_verified": "2026-09-25",
      "source": "https://developers.openai.com/api/docs/pricing",
      "unverified": false,
      "notes": "Batch/Flex: $1/$5. Fast mode: $4/$20. Long context: $4/$15."
    },
    {
      "id": "gpt-6-luna",
      "provider": "OpenAI",
      "model": "GPT-6 Luna",
      "input_per_1m": 0.1,
      "output_per_1m": 0.5,
      "last_verified": "2026-09-25",
      "source": "https://developers.openai.com/api/docs/pricing",
      "unverified": false,
      "notes": "Budget tier. Batch/Flex: $0.05/$0.25. Long context: $0.20/$0.75."
    },
    {
      "id": "gpt-5-6-sol",
      "provider": "OpenAI",
      "model": "GPT-5.6 Sol",
      "input_per_1m": 4.0,
      "output_per_1m": 20.0,
      "last_verified": "2026-09-25",
      "source": "https://developers.openai.com/api/docs/pricing",
      "unverified": false,
      "notes": "Promo price through at least Nov 21, 2026. Long context: $8/$30."
    },
    {
      "id": "gpt-5-6-cyber",
      "provider": "OpenAI",
      "model": "GPT-5.6 Cyber",
      "input_per_1m": 12.5,
      "output_per_1m": 75.0,
      "last_verified": "2026-09-25",
      "source": "https://developers.openai.com/api/docs/pricing",
      "unverified": false,
      "notes": "Cyber-specialized tier."
    },
    {
      "id": "gpt-5-3-codex",
      "provider": "OpenAI",
      "model": "GPT-5.3 Codex",
      "input_per_1m": 1.75,
      "output_per_1m": 14.0,
      "last_verified": "2026-09-25",
      "source": "https://developers.openai.com/api/docs/pricing",
      "unverified": false,
      "notes": "Codex coding model."
    },
    {
      "id": "gpt-5-4",
      "provider": "OpenAI",
      "model": "GPT-5.4",
      "input_per_1m": 2.5,
      "output_per_1m": 15.0,
      "last_verified": "2026-09-25",
      "source": "https://benchlm.ai/openai/api-pricing",
      "unverified": true,
      "notes": "Index-only figure; not present on OpenAI's live pricing page. Verify before relying."
    },
    {
      "id": "gpt-5-4-mini",
      "provider": "OpenAI",
      "model": "GPT-5.4 Mini",
      "input_per_1m": 0.75,
      "output_per_1m": 4.5,
      "last_verified": "2026-09-25",
      "source": "https://benchlm.ai/openai/api-pricing",
      "unverified": true,
      "notes": "Index-only figure; not present on OpenAI's live pricing page."
    },
    {
      "id": "gpt-5-4-nano",
      "provider": "OpenAI",
      "model": "GPT-5.4 Nano",
      "input_per_1m": 0.2,
      "output_per_1m": 1.25,
      "last_verified": "2026-09-25",
      "source": "https://benchlm.ai/openai/api-pricing",
      "unverified": true,
      "notes": "Index-only figure; not present on OpenAI's live pricing page."
    },
    {
      "id": "gpt-5-1",
      "provider": "OpenAI",
      "model": "GPT-5.1",
      "input_per_1m": 1.25,
      "output_per_1m": 10.0,
      "last_verified": "2026-09-25",
      "source": "https://www.requesty.ai/models/openai/gpt-5.1",
      "unverified": true,
      "notes": "Index-only figure; not present on OpenAI's live pricing page."
    },
    {
      "id": "gpt-5-nano",
      "provider": "OpenAI",
      "model": "GPT-5 Nano",
      "input_per_1m": 0.05,
      "output_per_1m": 0.4,
      "last_verified": "2026-09-25",
      "source": "https://anotherwrapper.com/tools/llm-pricing",
      "unverified": true,
      "notes": "Index-only figure; not present on OpenAI's live pricing page."
    },
    {
      "id": "claude-fable-5-1",
      "input_per_1m": 10.0,
      "last_verified": "2026-09-25",
      "model": "Claude Fable 5.1",
      "notes": "Most capable Anthropic tier; cache reads $0.25/M.",
      "output_per_1m": 50.0,
      "provider": "Anthropic",
      "source": "https://benchlm.ai/anthropic/api-pricing",
      "unverified": false
    },
    {
      "id": "claude-opus-5-5",
      "input_per_1m": 4.0,
      "last_verified": "2026-09-25",
      "model": "Claude Opus 5.5",
      "notes": "Released 2026-09-22. 20% below Opus 5; cache reads $0.20/M.",
      "output_per_1m": 20.0,
      "provider": "Anthropic",
      "source": "https://www.testingcatalog.com/anthropic-launches-claude-opus-5-5-with-lower-api-costs/",
      "unverified": false
    },
    {
      "id": "claude-opus-5",
      "input_per_1m": 5.0,
      "last_verified": "2026-09-25",
      "model": "Claude Opus 5",
      "notes": "",
      "output_per_1m": 25.0,
      "provider": "Anthropic",
      "source": "https://benchlm.ai/anthropic/api-pricing",
      "unverified": false
    },
    {
      "id": "claude-sonnet-5",
      "input_per_1m": 2.0,
      "last_verified": "2026-09-25",
      "model": "Claude Sonnet 5",
      "notes": "Introductory rate made permanent in September 2026.",
      "output_per_1m": 10.0,
      "provider": "Anthropic",
      "source": "https://benchlm.ai/anthropic/api-pricing",
      "unverified": false
    },
    {
      "id": "claude-haiku-4-5",
      "input_per_1m": 1.0,
      "last_verified": "2026-09-25",
      "model": "Claude Haiku 4.5",
      "notes": "Low-cost tier for high-volume tasks.",
      "output_per_1m": 5.0,
      "provider": "Anthropic",
      "source": "https://benchlm.ai/anthropic/api-pricing",
      "unverified": false
    },
    {
      "id": "gemini-3-1-pro-preview",
      "input_per_1m": 2.0,
      "last_verified": "2026-09-25",
      "model": "Gemini 3.1 Pro Preview",
      "notes": "Above 200K prompt tokens: $4 input / $18 output. Paid-only, no free tier.",
      "output_per_1m": 12.0,
      "provider": "Google",
      "source": "https://www.requesty.ai/models/google/gemini-3.1-pro-preview",
      "unverified": false
    },
    {
      "id": "gemini-3-8-flash",
      "input_per_1m": 0.75,
      "last_verified": "2026-09-25",
      "model": "Gemini 3.8 Flash",
      "notes": "Intro price through 2026-12-31, then $1.50/$7.50.",
      "output_per_1m": 3.75,
      "provider": "Google",
      "source": "https://innfactory.ai/en/ai-models/google-gemini/",
      "unverified": false
    },
    {
      "id": "gemini-3-5-flash",
      "input_per_1m": 1.5,
      "last_verified": "2026-09-25",
      "model": "Gemini 3.5 Flash",
      "notes": "",
      "output_per_1m": 9.0,
      "provider": "Google",
      "source": "https://costgoat.com/pricing/gemini-api",
      "unverified": false
    },
    {
      "id": "gemini-3-flash-preview",
      "input_per_1m": 0.5,
      "last_verified": "2026-09-25",
      "model": "Gemini 3 Flash Preview",
      "notes": "Mid-tier Flash, preview pricing.",
      "output_per_1m": 3.0,
      "provider": "Google",
      "source": "https://axis-intelligence.com/multimodal-ai-statistics/",
      "unverified": false
    },
    {
      "id": "gemini-3-5-flash-lite",
      "input_per_1m": 0.3,
      "last_verified": "2026-09-25",
      "model": "Gemini 3.5 Flash-Lite",
      "notes": "",
      "output_per_1m": 2.5,
      "provider": "Google",
      "source": "https://costgoat.com/pricing/gemini-api",
      "unverified": false
    },
    {
      "id": "gemini-3-1-flash-lite",
      "input_per_1m": 0.25,
      "last_verified": "2026-09-25",
      "model": "Gemini 3.1 Flash-Lite",
      "notes": "",
      "output_per_1m": 1.5,
      "provider": "Google",
      "source": "https://costgoat.com/pricing/gemini-api",
      "unverified": false
    },
    {
      "id": "gemini-2-5-flash-lite",
      "input_per_1m": 0.1,
      "last_verified": "2026-09-25",
      "model": "Gemini 2.5 Flash-Lite",
      "notes": "Cheapest Gemini entry point.",
      "output_per_1m": 0.4,
      "provider": "Google",
      "source": "https://anotherwrapper.com/tools/llm-pricing",
      "unverified": false
    },
    {
      "id": "grok-4-7",
      "input_per_1m": 2.0,
      "last_verified": "2026-09-25",
      "model": "Grok 4.7",
      "notes": "Released 2026-09-21. Cached input $0.50/M. Prompts of 200K+ tokens bill $4/$12 for the whole request.",
      "output_per_1m": 6.0,
      "provider": "xAI",
      "source": "https://tokenscost.com/blog/grok-4-7-pricing-context-window-benchmarks",
      "unverified": false
    },
    {
      "id": "grok-4-6",
      "input_per_1m": 2.0,
      "last_verified": "2026-09-25",
      "model": "Grok 4.6",
      "notes": "Cached input $0.50/M.",
      "output_per_1m": 6.0,
      "provider": "xAI",
      "source": "https://www.requesty.ai/models/xai/grok-4.6",
      "unverified": false
    },
    {
      "id": "grok-4-20",
      "input_per_1m": 1.25,
      "last_verified": "2026-09-25",
      "model": "Grok 4.20",
      "notes": "Value tier of the Grok 4.x family.",
      "output_per_1m": 2.5,
      "provider": "xAI",
      "source": "https://costgoat.com/pricing/grok-api",
      "unverified": false
    },
    {
      "id": "grok-build-0-1",
      "input_per_1m": 1.0,
      "last_verified": "2026-09-25",
      "model": "Grok Build 0.1",
      "notes": "Coding-focused model.",
      "output_per_1m": 2.0,
      "provider": "xAI",
      "source": "https://costgoat.com/pricing/grok-api",
      "unverified": false
    },
    {
      "id": "llama-3-1-8b-instant",
      "input_per_1m": 0.05,
      "last_verified": "2026-09-25",
      "model": "Llama 3.1 8B Instant",
      "notes": "Cheapest tracked model overall; second-hand citation of groq.com/pricing. Direct fetch of groq.com/pricing (2026-09-25) did not list these models or prices — page may be JS-rendered. Verify manually.",
      "output_per_1m": 0.08,
      "provider": "Groq",
      "source": "https://www.spheron.network/blog/groq-api-pricing-2026-cost-per-token-vs-gpu-rental/",
      "unverified": true
    },
    {
      "id": "llama-3-3-70b-versatile",
      "input_per_1m": 0.59,
      "last_verified": "2026-09-25",
      "model": "Llama 3.3 70B Versatile",
      "notes": "Second-hand citation of groq.com/pricing. Direct fetch of groq.com/pricing (2026-09-25) did not list these models or prices — page may be JS-rendered. Verify manually.",
      "output_per_1m": 0.79,
      "provider": "Groq",
      "source": "https://www.spheron.network/blog/groq-api-pricing-2026-cost-per-token-vs-gpu-rental/",
      "unverified": true
    },
    {
      "id": "gpt-oss-20b",
      "input_per_1m": 0.075,
      "last_verified": "2026-09-25",
      "model": "GPT-OSS 20B",
      "notes": "CONFLICT: index A reports $0.075/$0.30, index B reports $0.10/$0.50. Manual check of groq.com/pricing required.",
      "output_per_1m": 0.3,
      "provider": "Groq",
      "source": "https://www.requesty.ai/models/groq/openai-gpt-oss-20b",
      "unverified": true
    },
    {
      "id": "gpt-oss-120b",
      "input_per_1m": 0.15,
      "last_verified": "2026-09-25",
      "model": "GPT-OSS 120B",
      "notes": "CONFLICT: index A reports $0.15/$0.60 output, index B reports $0.15/$0.75 output. Manual check required.",
      "output_per_1m": 0.6,
      "provider": "Groq",
      "source": "https://www.requesty.ai/models/groq/openai-gpt-oss-120b",
      "unverified": true
    },
    {
      "id": "deepseek-v4-1-flash",
      "input_per_1m": 0.15,
      "last_verified": "2026-09-25",
      "model": "DeepSeek V4.1 Flash",
      "notes": "Off-peak rates shown (most hours). Peak (01:00-04:00, 06:00-10:00 UTC, Mon-Fri) doubles to $0.30/$1.20. Cache hits ~2% of input rate.",
      "output_per_1m": 0.6,
      "provider": "DeepSeek",
      "source": "https://deepseek.ai/pricing",
      "unverified": false
    },
    {
      "id": "deepseek-v4-pro",
      "input_per_1m": 0.66,
      "last_verified": "2026-09-25",
      "model": "DeepSeek V4 Pro",
      "notes": "Off-peak rates shown. Peak doubles to $1.32/$3.96.",
      "output_per_1m": 1.98,
      "provider": "DeepSeek",
      "source": "https://deepseek.ai/pricing",
      "unverified": false
    },
    {
      "id": "mistral-large-3",
      "input_per_1m": 0.5,
      "last_verified": "2026-09-25",
      "model": "Mistral Large 3",
      "notes": "CONFLICT: one Sep 2026 index reports $0.50/$1.50, another reports $2.00/$5.00 citing Mistral docs. Manual check of docs.mistral.ai/platform/pricing required.",
      "output_per_1m": 1.5,
      "provider": "Mistral",
      "source": "https://www.aipricing.guru/mistral-pricing/",
      "unverified": true
    },
    {
      "id": "mistral-medium-3-5",
      "input_per_1m": 1.5,
      "last_verified": "2026-09-25",
      "model": "Mistral Medium 3.5",
      "notes": "Single index source; flagship-line conflict casts doubt. Verify against Mistral docs.",
      "output_per_1m": 7.5,
      "provider": "Mistral",
      "source": "https://www.aipricing.guru/mistral-pricing/",
      "unverified": true
    },
    {
      "id": "mistral-small-4",
      "input_per_1m": 0.15,
      "last_verified": "2026-09-25",
      "model": "Mistral Small 4",
      "notes": "Single index source. Verify against Mistral docs.",
      "output_per_1m": 0.6,
      "provider": "Mistral",
      "source": "https://www.aipricing.guru/mistral-pricing/",
      "unverified": true
    },
    {
      "id": "codestral",
      "input_per_1m": 0.3,
      "last_verified": "2026-09-25",
      "model": "Codestral",
      "notes": "Single index source. Verify against Mistral docs.",
      "output_per_1m": 0.9,
      "provider": "Mistral",
      "source": "https://www.aipricing.guru/mistral-pricing/",
      "unverified": true
    }
  ],
  "price_moves": [
    {
      "date": "2026-09-22",
      "detail": "Anthropic's new Opus tier undercuts Opus 5's $5/$25 by 20% and cuts cache reads to $0.20/M — the cheapest per-token Anthropic has ever sold a frontier coding model.",
      "headline": "Claude Opus 5.5 launched at $4/$20 per 1M tokens",
      "source": "https://www.testingcatalog.com/anthropic-launches-claude-opus-5-5-with-lower-api-costs/"
    },
    {
      "date": "2026-09-21",
      "detail": "xAI shipped a larger base model without moving the rate card — an unusually still price in a year of weekly cuts. Cached input stays $0.50/M; 200K+ token prompts bill the higher $4/$12 tier.",
      "headline": "Grok 4.7 launched at the same $2/$6 as Grok 4.6",
      "source": "https://aitechconnect.in/news/grok-4-7-same-price-terminal-bench-doubles-2026"
    },
    {
      "date": "2026-09-09",
      "detail": "The cheapest frontier-adjacent API now splits the day: $0.15/$0.60 off-peak, $0.30/$1.20 during UTC peak hours (01:00-04:00 and 06:00-10:00, Mon-Fri).",
      "headline": "DeepSeek V4.1 Flash went GA with peak/off-peak billing",
      "source": "https://deepseek.ai/pricing"
    },
    {
      "date": "2026-09-01",
      "detail": "What was a launch promo through August became the standing price, putting Sonnet 5 at parity with GPT-5.6 Terra on input and under it on output.",
      "headline": "Anthropic made Sonnet 5's $2/$10 introductory rate permanent",
      "source": "https://benchlm.ai/anthropic/api-pricing"
    },
    {
      "date": "2026-09-16",
      "detail": "Google kept the $0.75/$3.75 rate through Dec 31, 2026, before it steps up to $1.50/$7.50 in January — plan 2027 budgets accordingly.",
      "headline": "Gemini 3.8 Flash held at intro pricing through year-end",
      "source": "https://innfactory.ai/en/ai-models/google-gemini/"
    },
    {
      "date": "2026-09-25",
      "headline": "GPT-6 family took the flagship slot: Astra $10/$50, Sol $2/$10, Luna $0.10/$0.50",
      "detail": "OpenAI's live pricing page now leads with GPT-6. Astra at $10/$50 doubles the old 5.6 Sol list price; Sol at $2/$10 and Luna at $0.10/$0.50 keep the mid and budget lanes. 5.6 Sol holds its $4/$20 promo through Nov 21.",
      "source": "https://developers.openai.com/api/docs/pricing"
    }
  ]
}