openclaw.json

// In agents.models:
"nvidia/moonshotai/kimi-k2.5": {
  "alias": "kimi"
}

// In the root
"models": {
  "mode": "merge",
  "providers": {
    "nvidia": {
      "baseUrl": "https://integrate.api.nvidia.com/v1",
      "apiKey": "<API_KEY>",
      "api": "openai-completions",
      "models": [
        {
          "id": "moonshotai/kimi-k2.5",
          "name": "Kimi K2.5",
          "reasoning": true,
          "input": [
            "text",
            "image"
          ],
          "cost": {
            "input": 0,
            "output": 0,
            "cacheRead": 0,
            "cacheWrite": 0
          },
          "contextWindow": 256000,
          "maxTokens": 8192
        }
      ]
    }
  }
}

Try guys kimi k2.5 through ollama. It is much faster.

models.providers.

"ollama": {
"baseUrl": "http://127.0.0.1:11434/v1",
"apiKey": "ollama",
"api": "openai-responses",
"models": [
{
"id": "kimi-k2.5:cloud",
"name": "Kimi K2.5 Cloud",
"reasoning": true,
"input": ["text", "image"],
"cost": { "input": 0, "output": 0, "cacheRead": 0, "cacheWrite": 0 },
"contextWindow": 262144,
"maxTokens": 2621440
},
]
},

agents.defaults.models.

    "nvidia/moonshotai/kimi-k2.5": {
      "alias": "nvidia"
    },
    "ollama/kimi-k2.5:cloud": {
      "alias": "kimi"

Where do you put the API key, replace “Ollama”? Also is it free ? The nvidia one is free.

…

On Sat, Feb 21, 2026 at 2:07 AM N.O.V.A. ***@***.***> wrote: ***@***.**** commented on this gist. ------------------------------ Try guys kimi k2.5 through ollama. It is much faster. models.providers. "ollama": { "baseUrl": "http://127.0.0.1:11434/v1", "apiKey": "ollama", "api": "openai-responses", "models": [ { "id": "kimi-k2.5:cloud", "name": "Kimi K2.5 Cloud", "reasoning": true, "input": ["text", "image"], "cost": { "input": 0, "output": 0, "cacheRead": 0, "cacheWrite": 0 }, "contextWindow": 262144, "maxTokens": 2621440 }, ] }, agents.defaults.models. "nvidia/moonshotai/kimi-k2.5": { "alias": "nvidia" }, "ollama/kimi-k2.5:cloud": { "alias": "kimi" — Reply to this email directly, view it on GitHub <https://gist.github.com/haltakov/72f732bacb7c81a056fc1853cc6e970a#gistcomment-5996452> or unsubscribe <https://github.com/notifications/unsubscribe-auth/B6NRVXSTTPCA2QQPY2JLBU34M774TBFKMF2HI4TJMJ2XIZLTSOBKK5TBNR2WLKJSGYYDENBSGE4TJJDOMFWWLKDBMN2G64S7NFSIFJLWMFWHKZNEORZHKZNENZQW2ZN3ORUHEZLBMRPXAYLSORUWG2LQMFXHIX3BMN2GS5TJOR4YFJLWMFWHKZNEM5UXG5FENZQW2ZNLORUHEZLBMRPXI6LQMWWHG5LCNJSWG5C7OR4XAZNLI5UXG5CDN5WW2ZLOOSTHI33QNFRXHEMCUR2HS4DFURTWS43UUV3GC3DVMWUTCNBVGMZTKOBVGKTXI4TJM5TWK4VGMNZGKYLUMU> . You are receiving this email because you commented on the thread. Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub> .

Where do you put the API key, replace “Ollama”? Also is it free ? The nvidia one is free.

Kimi K2.5 Cloud is currently free in Ollama, but you need an Ollama online account 😉

      "ollama": {
        "baseUrl": "http://localhost:11434/v1",
        "apiKey": "ollama-local",
        "api": "openai-completions",
        "models": [
          {
            "id": "kimi-k2.5:cloud",
            "name": "Kimi K2.5 Cloud",
            "cost": {
              "input": 0.002,
              "output": 0.006,
              "cacheRead": 0,
              "cacheWrite": 0
            },
            "contextWindow": 262144,
            "maxTokens": 262144
          }
        ]
      }

Yes! Sry. I also messed up the output: 262144.
Free with Ollama free account. I am not sure about the limit cos I didn't follow tokens, but I hit the limit around the same on bith nvidia and ollama. But ollama is much faster. At least for me

Yes! Sry. I also messed up the output: 262144. Free with Ollama free account. I am not sure about the limit cos I didn't follow tokens, but I hit the limit around the same on bith nvidia and ollama. But ollama is much faster. At least for me

Here are the essential facts regarding Kimi-K2.5:cloud quotas on the Ollama Cloud Free Tier:

• Token Context: Supports up to 256,000 (256K) tokens for input.
• Generation Limit: Output is capped at approximately 16,384 tokens per response.
• Usage Model: Operates on a percentage-based system (usage credits) rather than a fixed token count.
• Hourly Quota: Roughly 250,000 input tokens per hour (subject to system load).
• Request Rate: Approximately 135 requests every 5 hours for typical chat usage.
• Resource Intensity: As a 1-trillion parameter model, Kimi-K2.5 consumes your usage percentage faster than smaller models (like Llama 3).
• Monitoring: Live usage tracking is available in your Ollama Dashboard under account settings.

My config: https://github.com/rainman74/OpenClaw-Configs/blob/main/.openclaw/Windows/openclaw.json

The openclaw.json haltakov posted still rocks now!
Thank you!