{
  "key": "kimi-k3-flex",
  "name": "Kimi K3 Flex",
  "description": "Kimi K3 on the flex tier: discounted, best-effort latency, requests may be held under load",
  "lab": "moonshotai",
  "family": "kimi-k3",
  "release_date": "2026-07-16",
  "knowledge": null,
  "open_weights": true,
  "has_free_offering": false,
  "published": true,
  "modalities": {
    "input": [
      "text",
      "image"
    ],
    "output": [
      "text"
    ]
  },
  "capabilities": {
    "reasoning": true,
    "tool_call": true,
    "structured_output": true,
    "attachment": true,
    "reasoning_options": [
      {
        "type": "effort",
        "values": [
          "none",
          "low",
          "high",
          "max"
        ]
      },
      {
        "type": "budget_tokens"
      }
    ]
  },
  "limit": {
    "context": 1048560,
    "output": 1048560
  },
  "offerings": [
    {
      "provider": "neuralwatt",
      "provider_model_id": "kimi-k3-flex",
      "name": "Kimi K3 Flex",
      "variants": [],
      "free": false,
      "priced": true,
      "cost": {
        "input": 1.95,
        "output": 9.75,
        "cache_read": 0.195,
        "cache_write": null,
        "reasoning": null,
        "context_over_200k": null,
        "tiers": null
      },
      "limit": {
        "context": 1048560,
        "output": 1048560,
        "input": null
      },
      "status": null,
      "last_updated": "2026-07-16"
    }
  ]
}