Baltor Get started

Models

Kimi-K3

By moonshotai. 2.78 trillion parameters, a context window of 1,048,576 tokens and the licence other.

Facts

Released
2026-06-13 the day the repository was first published on Hugging Face Hugging Face, read
Licence
other the licence the model card declares Hugging Face, read
Open weights
Yes the weights are published in this Hugging Face repository Hugging Face, read
Parameters
2.78 trillion counted from the safetensors weight files Hugging Face, read
Active parameters
Unknown
Knowledge cutoff
Unknown
Context window
  • 1,048,576 tokens the maximum position embeddings in the model configuration model configuration, read
Longest output
Unknown
Tool calling
Unknown
Structured output
Unknown
Reasoning controls
Unknown
Inputs and outputs
text, image in, text out Hugging Face, read
Good for
  • Vision: Hugging Face files it under image-text-to-text Hugging Face, read

Prices

Each provider's own price in US dollars per million tokens, as an openly licensed source lists it
ProviderInputOutputContextAs ofSource
Nvidia0 USD0 USD1,048,576 older than 30 days; check the providermodels.dev, read
OpenRouter0.239 USD11.94 USD1,048,576 older than 30 days; check the providermodels.dev, read
NanoGPT1.65 USD8.25 USD1,048,576 older than 30 days; check the providermodels.dev, read
IteraCompute2.39 USD11.99 USD1,048,576 older than 30 days; check the providermodels.dev, read
sail2.50 USD12.50 USD1,048,576LiteLLM, read
SiliconFlow2.70 USD13.50 USD1,048,576 older than 30 days; check the providermodels.dev, read
Wallaby2.70 USD13.50 USD1,048,576 older than 30 days; check the providermodels.dev, read
Deep Infra2.85 USD14.25 USD1,048,576 older than 30 days; check the providermodels.dev, read
ai&3.00 USD12.50 USD1,048,576 older than 30 days; check the providermodels.dev, read
Abacus3.00 USD15.00 USD1,048,576 older than 30 days; check the providermodels.dev, read
Arcee3.00 USD15.00 USD1,048,576 older than 30 days; check the providermodels.dev, read
Baseten3.00 USD15.00 USD1,048,576 older than 30 days; check the providermodels.dev, read
Berget.AI3.00 USD15.00 USD327,680 older than 30 days; check the providermodels.dev, read
Cloudflare AI Gateway3.00 USD15.00 USD1,048,576 older than 30 days; check the providermodels.dev, read
CommandCode3.00 USD15.00 USD1,000,000 older than 30 days; check the providermodels.dev, read
Hugging Face3.00 USD15.00 USD1,000,000 older than 30 days; check the providermodels.dev, read
Impossibl3.00 USD15.00 USD1,048,576 older than 30 days; check the providermodels.dev, read
Kilo Gateway3.00 USD15.00 USD1,048,576 older than 30 days; check the providermodels.dev, read
Modal3.00 USD15.00 USD1,048,576 older than 30 days; check the providermodels.dev, read
Nebius Token Factory3.00 USD15.00 USD1,048,576 older than 30 days; check the providermodels.dev, read
Novita AI3.00 USD15.00 USD1,048,576 older than 30 days; check the providermodels.dev, read
Ofox3.00 USD15.00 USD1,048,576 older than 30 days; check the providermodels.dev, read
Pioneer3.00 USD15.00 USD1,000,000 older than 30 days; check the providermodels.dev, read
Standard Compute3.00 USD15.00 USD1,048,576 older than 30 days; check the providermodels.dev, read
TensorX3.00 USD15.00 USD1,048,576 older than 30 days; check the providermodels.dev, read
Together AI3.00 USD15.00 USD1,048,576 older than 30 days; check the providermodels.dev, read
TokenGo3.00 USD15.00 USD1,048,576 older than 30 days; check the providermodels.dev, read
Vercel AI Gateway3.00 USD15.00 USD1,000,000 older than 30 days; check the providermodels.dev, read
ZenMux3.00 USD15.00 USD1,048,576 older than 30 days; check the providermodels.dev, read
NEAR AI Cloud3.30 USD16.50 USD1,048,576 older than 30 days; check the providermodels.dev, read
Prime Inference3.45 USD17.25 USD1,048,576 older than 30 days; check the providermodels.dev, read

Run it on your own hardware

Weights are the sizes of the files a source lists, or an estimate from the parameter count where none does. Memory is an estimate: weights plus KV cache plus 512 MiB and 5 percent of the weights for runtime buffers. Check it against your hardware.

QuantizationWeightsMemory at 8,192 tokensMemory at 32,768 tokens
UD-TQ1_0474 GiBUnknown: no KV cache numbersUnknown: no KV cache numbers
UD-IQ1_S553 GiBUnknown: no KV cache numbersUnknown: no KV cache numbers
UD-IQ1_M604 GiBUnknown: no KV cache numbersUnknown: no KV cache numbers
UD-IQ2_XXS662 GiBUnknown: no KV cache numbersUnknown: no KV cache numbers
UD-Q2_K_XL802 GiBUnknown: no KV cache numbersUnknown: no KV cache numbers
UD-Q4_K_XL1,405 GiBUnknown: no KV cache numbersUnknown: no KV cache numbers

Start it with a local runtime

  • Ollama
    ollama run hf.co/unsloth/Kimi-K3-GGUF:UD-TQ1_0
  • LM Studio
    lms get unsloth/Kimi-K3-GGUF
    lms server start
  • llama.cpp server
    llama-server -hf unsloth/Kimi-K3-GGUF:UD-TQ1_0 --jinja
  • vLLM
    vllm serve moonshotai/Kimi-K3
  • SGLang
    sglang serve --model-path moonshotai/Kimi-K3 --port 30000

Use it from your harness

Set up for Abacus with the model moonshotai/Kimi-K3. Each endpoint page has the same setup for its own address.

OpenCode

Put this in opencode.json in your project folder:

{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "abacus": {
      "npm": "@ai-sdk/openai-compatible",
      "name": "Abacus",
      "options": {
        "baseURL": "https://routellm.abacus.ai/v1",
        "apiKey": "{env:ABACUS_API_KEY}"
      },
      "models": {
        "moonshotai/Kimi-K3": {
          "name": "Kimi-K3"
        }
      }
    }
  }
}
  • OpenCode reads any OpenAI-compatible address through the @ai-sdk/openai-compatible package, and an address that speaks the Responses API through @ai-sdk/openai.

From OpenCode documentation, read .

Pi

Put this in ~/.pi/agent/models.json:

{
  "providers": {
    "abacus": {
      "baseUrl": "https://routellm.abacus.ai/v1",
      "api": "openai-completions",
      "apiKey": "$ABACUS_API_KEY",
      "models": [
        {
          "id": "moonshotai/Kimi-K3"
        }
      ]
    }
  }
}
  • The apiKey field can name an environment variable as $NAME.

From Pi documentation, read .

Codex

Codex speaks only the Responses API, and Abacus documents no Responses address. A gateway that offers one can sit in between.

  • Codex speaks the Responses API only: responses is the one supported wire API of a custom provider. Ollama and LM Studio are built in and start with --oss.

From Codex documentation, read .

Claude Code

Claude Code sends Anthropic Messages requests, and Abacus documents no such address. A gateway that translates to that API can sit in between.

  • Claude Code sends Anthropic Messages requests to ANTHROPIC_BASE_URL. Anthropic says it does not support routing Claude Code to models other than Claude through any gateway, so some features may not work with another model.

From Claude Code documentation, read .

Other ways to reach it

Published results

These are results other people published. Baltor did not run them and does not rank models by them.

Two more publishers are linked, never copied, because their terms do not allow it:

  • Artificial Analysis publishes its own intelligence, speed and price measurements.
  • OpenRouter lists current prices and uptime for each provider it routes to.