{
 "lastUpdated": "2026-10-15",
 "currency": "USD per 1M tokens (standard, short-context, list price)",
 "sources": [
  {
   "id": "openai",
   "name": "OpenAI rate card",
   "url": "https://help.openai.com/en/articles/20001415-chatgpt-rate-card-enterprise-token-based-pricing"
  },
  {
   "id": "anthropic",
   "name": "Anthropic list prices (eff. 2026-10-07)",
   "url": "https://assets.claude.com/1ef7a4a8e373004bc89d710b0a6f3f219323344f.pdf"
  },
  {
   "id": "deepseek",
   "name": "DeepSeek API pricing",
   "url": "https://api-docs.deepseek.com/quick_start/pricing/"
  },
  {
   "id": "promptcost",
   "name": "PromptCost LLM API pricing (Oct 2026)",
   "url": "https://promptcost.dev/llm-api-pricing"
  },
  {
   "id": "intuition",
   "name": "IntuitionLabs pricing comparison (checked 2026-10-10)",
   "url": "https://intuitionlabs.ai/articles/llm-api-pricing-comparison-2025"
  },
  {
   "id": "alibaba",
   "name": "Alibaba Cloud Model Studio pricing (updated 2026-10-13)",
   "url": "https://www.alibabacloud.com/help/en/model-studio/model-pricing"
  },
  {
   "id": "wallaby",
   "name": "Wallaby open-weight pricing snapshot (2026-10-04)",
   "url": "https://wallabytoken.com/blog/p/open-weight-llm-api-pricing"
  },
  {
   "id": "minimax",
   "name": "MiniMax API pricing",
   "url": "https://platform.minimax.io/docs/pricing/overview"
  },
  {
   "id": "modelpick",
   "name": "modelpick Chinese model pricing (Oct 2026)",
   "url": "https://modelpick.dev/chinese-ai-models/"
  },
  {
   "id": "google",
   "name": "Gemini Developer API pricing (checked 2026-10-15)",
   "url": "https://ai.google.dev/gemini-api/docs/pricing"
  }
 ],
 "models": [
  {
   "name": "GPT-6 Astra",
   "provider": "OpenAI",
   "in": 10,
   "out": 50,
   "ctx": "Unknown",
   "open": false,
   "cats": [
    "reasoning"
   ],
   "strengths": "Top-tier OpenAI pro model",
   "use": "Hard reasoning, research, high-stakes analysis",
   "src": "openai",
   "country": "US",
   "tier": 5,
   "uses": [
    "agents",
    "chat",
    "coding",
    "reasoning"
   ]
  },
  {
   "name": "Claude Fable 5.1",
   "provider": "Anthropic",
   "in": 10,
   "out": 50,
   "ctx": "1M",
   "open": false,
   "cats": [
    "reasoning",
    "writing"
   ],
   "strengths": "Anthropic's top model",
   "use": "Complex agents, long documents, deep analysis",
   "src": "intuition",
   "country": "US",
   "tier": 5,
   "uses": [
    "agents",
    "coding",
    "reasoning",
    "writing"
   ]
  },
  {
   "name": "Claude Opus 5.5",
   "provider": "Anthropic",
   "in": 4,
   "out": 20,
   "ctx": "1M",
   "open": false,
   "cats": [
    "coding",
    "writing",
    "reasoning"
   ],
   "strengths": "Strong coding & agentic work",
   "use": "Vibe coding, refactors, long-form writing",
   "src": "anthropic",
   "country": "US",
   "tier": 5,
   "uses": [
    "agents",
    "coding",
    "reasoning",
    "writing"
   ]
  },
  {
   "name": "Gemini 4 Argon",
   "provider": "Google",
   "in": 2,
   "out": 10,
   "ctx": "1M",
   "open": false,
   "cats": [
    "reasoning",
    "coding"
   ],
   "strengths": "Google flagship; intro price matches Sol/Sonnet until Nov 30",
   "use": "Hard reasoning, coding, long agentic work",
   "src": "google",
   "note": "Introductory $2/$10 through 2026-11-30; then $4/$20 from Dec 1",
   "country": "US",
   "tier": 5,
   "uses": [
    "agents",
    "chat",
    "coding",
    "multimodal",
    "reasoning"
   ]
  },
  {
   "name": "GPT-6.1 Sol",
   "provider": "OpenAI",
   "in": 2,
   "out": 10,
   "ctx": "Unknown",
   "open": false,
   "cats": [
    "coding",
    "reasoning"
   ],
   "strengths": "Flagship value, cheap cached input ($0.10)",
   "use": "General assistant, coding, apps",
   "src": "openai",
   "country": "US",
   "tier": 4,
   "uses": [
    "agents",
    "chat",
    "coding",
    "multimodal",
    "reasoning"
   ]
  },
  {
   "name": "Claude Sonnet 5.5",
   "provider": "Anthropic",
   "in": 2,
   "out": 10,
   "ctx": "1M",
   "open": false,
   "cats": [
    "coding",
    "writing"
   ],
   "strengths": "Balanced cost/quality",
   "use": "Everyday coding, writing, agents",
   "src": "anthropic",
   "country": "US",
   "tier": 4,
   "uses": [
    "agents",
    "chat",
    "coding",
    "writing"
   ]
  },
  {
   "name": "Grok 4.7",
   "provider": "xAI",
   "in": 2,
   "out": 6,
   "ctx": "500K",
   "open": false,
   "cats": [
    "reasoning",
    "coding"
   ],
   "strengths": "xAI's most capable; low output price",
   "use": "Reasoning, real-time-aware apps, coding",
   "src": "promptcost",
   "country": "US",
   "tier": 4,
   "uses": [
    "chat",
    "coding",
    "reasoning"
   ]
  },
  {
   "name": "GPT-5.3-Codex",
   "provider": "OpenAI",
   "in": 1.75,
   "out": 14,
   "ctx": "Unknown",
   "open": false,
   "cats": [
    "coding"
   ],
   "strengths": "Coding-specialized",
   "use": "Code generation & agents",
   "src": "openai",
   "country": "US",
   "tier": 4,
   "uses": [
    "agents",
    "coding"
   ]
  },
  {
   "name": "DeepSeek V4 Pro",
   "provider": "DeepSeek",
   "in": 1.32,
   "out": 3.96,
   "ctx": "1M",
   "open": true,
   "cats": [
    "coding",
    "reasoning",
    "cheap",
    "open"
   ],
   "strengths": "Low price; 50% off-peak",
   "use": "Budget reasoning & coding",
   "src": "deepseek",
   "note": "Peak cache-miss rate; open-weights status: verify",
   "country": "CN",
   "tier": 4,
   "uses": [
    "coding",
    "reasoning"
   ]
  },
  {
   "name": "Grok 4.3",
   "provider": "xAI",
   "in": 1.25,
   "out": 2.5,
   "ctx": "1M",
   "open": false,
   "cats": [
    "cheap"
   ],
   "strengths": "Cheap, 1M context",
   "use": "High-volume chat, summarization",
   "src": "promptcost",
   "country": "US",
   "tier": 3,
   "uses": [
    "bulk",
    "chat",
    "writing"
   ]
  },
  {
   "name": "Gemini 3.8 Flash",
   "provider": "Google",
   "in": 0.75,
   "out": 3.75,
   "ctx": "1M",
   "open": false,
   "cats": [
    "cheap",
    "coding"
   ],
   "strengths": "Fast; promo price until 2026-12-31 (then $1.50/$7.50)",
   "use": "Fast apps, multimodal",
   "src": "intuition",
   "country": "US",
   "tier": 3,
   "uses": [
    "bulk",
    "chat",
    "coding",
    "multimodal"
   ]
  },
  {
   "name": "DeepSeek V4.1 Flash",
   "provider": "DeepSeek",
   "in": 0.3,
   "out": 1.2,
   "ctx": "1M",
   "open": true,
   "cats": [
    "cheap",
    "open"
   ],
   "strengths": "Very cheap; half off-peak",
   "use": "Bulk processing, bots",
   "src": "deepseek",
   "note": "Open-weights status: verify",
   "country": "CN",
   "tier": 3,
   "uses": [
    "agents",
    "bulk",
    "chat"
   ]
  },
  {
   "name": "Gemini 3.5 Flash-Lite",
   "provider": "Google",
   "in": 0.3,
   "out": 2.5,
   "ctx": "1M",
   "open": false,
   "cats": [
    "cheap"
   ],
   "strengths": "Low-latency",
   "use": "Classification, extraction",
   "src": "promptcost",
   "country": "US",
   "tier": 2,
   "uses": [
    "bulk",
    "multimodal"
   ]
  },
  {
   "name": "Mistral Small 4",
   "provider": "Mistral",
   "in": 0.15,
   "out": 0.6,
   "ctx": "262K",
   "open": true,
   "cats": [
    "cheap",
    "open"
   ],
   "strengths": "Cheap, efficient",
   "use": "Lightweight tasks, self-hosting",
   "src": "promptcost",
   "note": "Open-weights status: verify",
   "country": "EU",
   "tier": 2,
   "uses": [
    "bulk",
    "chat"
   ]
  },
  {
   "name": "gpt-oss-120B",
   "provider": "OpenAI (via Together)",
   "in": 0.15,
   "out": 0.6,
   "ctx": "131K",
   "open": true,
   "cats": [
    "cheap",
    "open"
   ],
   "strengths": "Open-weights OpenAI model",
   "use": "Self-host, private deployments",
   "src": "promptcost",
   "country": "US",
   "tier": 3,
   "uses": [
    "agents",
    "bulk",
    "chat"
   ]
  },
  {
   "name": "Claude Haiku 5.5",
   "provider": "Anthropic",
   "in": 0.1,
   "out": 0.5,
   "ctx": "1M",
   "open": false,
   "cats": [
    "cheap",
    "writing"
   ],
   "strengths": "Cheapest Claude (≤100K prompt; $0.50/$2.50 above)",
   "use": "High-volume support, routing",
   "src": "anthropic",
   "country": "US",
   "tier": 3,
   "uses": [
    "agents",
    "bulk",
    "chat",
    "writing"
   ]
  },
  {
   "name": "GPT-6 Luna",
   "provider": "OpenAI",
   "in": 0.1,
   "out": 0.5,
   "ctx": "1M",
   "open": false,
   "cats": [
    "cheap"
   ],
   "strengths": "Lowest-price flagship family tier",
   "use": "Bulk tasks, bots, classification",
   "src": "openai",
   "country": "US",
   "tier": 3,
   "uses": [
    "bulk",
    "chat"
   ]
  },
  {
   "name": "Kimi K3",
   "provider": "Moonshot AI",
   "country": "CN",
   "in": 3,
   "out": 15,
   "ctx": "1M",
   "open": true,
   "cats": [
    "reasoning",
    "coding",
    "open"
   ],
   "strengths": "China's top-priced reasoning model; open-weight",
   "use": "Agents, coding, long context",
   "src": "wallaby",
   "tier": 5,
   "uses": [
    "agents",
    "coding",
    "reasoning"
   ]
  },
  {
   "name": "Qwen3.8-Max",
   "provider": "Alibaba (Qwen)",
   "country": "CN",
   "in": 2,
   "out": 6,
   "ctx": "1M",
   "open": false,
   "cats": [
    "reasoning",
    "coding"
   ],
   "strengths": "Alibaba flagship; thinking + non-thinking modes",
   "use": "General reasoning, multilingual",
   "src": "alibaba",
   "note": "International deployment price",
   "tier": 4,
   "uses": [
    "chat",
    "reasoning",
    "writing"
   ]
  },
  {
   "name": "GLM-5.3",
   "provider": "Zhipu / Z.ai",
   "country": "CN",
   "in": 1.4,
   "out": 4.4,
   "ctx": "1M",
   "open": true,
   "cats": [
    "coding",
    "open"
   ],
   "strengths": "Strong value open-weight model",
   "use": "Coding agents, tool use",
   "src": "wallaby",
   "tier": 4,
   "uses": [
    "agents",
    "coding"
   ]
  },
  {
   "name": "MiniMax-M3",
   "provider": "MiniMax",
   "country": "CN",
   "in": 0.3,
   "out": 1.2,
   "ctx": "1M",
   "open": true,
   "cats": [
    "cheap",
    "open"
   ],
   "strengths": "Very cheap (permanent 50% off; 2x above 512K input)",
   "use": "Bulk agents, long context on a budget",
   "src": "minimax",
   "note": "Open-weights status: verify",
   "tier": 3,
   "uses": [
    "agents",
    "bulk"
   ]
  },
  {
   "name": "Qwen3.8 Flash",
   "provider": "Alibaba (Qwen)",
   "country": "CN",
   "in": 0.15,
   "out": 0.47,
   "ctx": "1M",
   "open": false,
   "cats": [
    "cheap"
   ],
   "strengths": "Ultra-low price",
   "use": "Classification, chat at scale",
   "src": "modelpick",
   "tier": 2,
   "uses": [
    "bulk",
    "chat"
   ]
  },
  {
   "name": "GLM-5.3-Flash",
   "provider": "Zhipu / Z.ai",
   "country": "CN",
   "in": 0.15,
   "out": 0.5,
   "ctx": "1M",
   "open": true,
   "cats": [
    "cheap",
    "open"
   ],
   "strengths": "Cheapest GLM tier",
   "use": "High-volume simple tasks",
   "src": "wallaby",
   "note": "Open-weights status: verify",
   "tier": 2,
   "uses": [
    "bulk",
    "chat"
   ]
  }
 ],
 "editorial": {
  "note": "Editorial picks by AIRaceBoard based on capability tier and price; not a benchmark score.",
  "overall": "Claude Fable 5.1",
  "us": "GPT-6 Astra",
  "cn": "Kimi K3",
  "eu": "Mistral Small 4",
  "open": "Kimi K3",
  "coding": "Claude Opus 5.5",
  "cheapGood": "DeepSeek V4.1 Flash",
  "changes": [
   "Gemini 4 Argon GA (Oct 13): intro $2/$10 per 1M through Nov 30, then $4/$20 (Google API pricing, checked Oct 15).",
   "China labs still quiet on new frontier drops since late August; US labs (Google Argon, Meta Muse Wonder) drove the week's news.",
   "Anthropic list prices updated Oct 7: Opus 5.5 $4/$20, Sonnet 5.5 $2/$10, Haiku 5.5 $0.10/$0.50.",
   "OpenAI GPT-6 family is current: Astra $10/$50, 6.1 Sol $2/$10, Luna $0.10/$0.50.",
   "Alibaba Qwen3.8-Max at $2/$6 (docs updated Oct 13).",
   "Kimi K3 output price reported moving to $15 per 1M (aiarch.dev changelog, Oct 14)."
  ],
  "vibe": {
   "pick": "Claude Opus 5.5",
   "agentic": "Claude Opus 5.5",
   "cheapIter": "GLM-5.3",
   "longCtx": "DeepSeek V4 Pro",
   "budget": "DeepSeek V4.1 Flash"
  }
 },
 "tierNote": "Tier = AIRaceBoard editorial capability tier (5 = frontier, 1 = basic), based on each lab's positioning and our review. Not a benchmark score."
}