moonshotai/kimi-k2.6 🔢📝🖼️❓ → 📝

⭐ Official ▶️ 9.7K runs 📅 Apr 2026 ⚙️ Cog 0.18.0 ⚖️ License
code-generation code-understanding image-to-text question-answering text-generation

About

Moonshot AI's frontier open model, built for long-horizon coding, agent swarms, and autonomous software engineering. 1 trillion parameters, 262k context window, vision and tool use.

Example Output

Prompt:

"Write a Python function group_anagrams(words) that groups strings by their anagram equivalence class. Include a concise docstring, type hints, and a short example in an if __name__ == '__main__': block. Output just the code, no prose."

Output

from collections import defaultdict
from typing import List, Dict


def group_anagrams(words: List[str]) -> Dict[str, List[str]]:
    """Group words by their anagram equivalence class.

    Returns a dictionary mapping sorted characters to the list of words
    that are anagrams of each other.
    """
    groups = defaultdict(list)
    for word in words:
        key = "".join(sorted(word))
        groups[key].append(word)
    return dict(groups)


if __name__ == '__main__':
    words = ["eat", "tea", "tan", "ate", "nat", "bat"]
    result = group_anagrams(words)
    print(result)
    # Example output: {'aet': ['eat', 'tea', 'ate'], 'ant': ['tan', 'nat'], 'abt': ['bat']}

Performance Metrics

3.88s Prediction Time
3.89s Total Time
All Input Parameters
{
  "prompt": "Write a Python function `group_anagrams(words)` that groups strings by their anagram equivalence class. Include a concise docstring, type hints, and a short example in an `if __name__ == '__main__':` block. Output just the code, no prose.",
  "max_tokens": 800,
  "reasoning_effort": "none"
}
Input Parameters
top_p Type: numberDefault: 1Range: 0 - 1
Nucleus sampling probability threshold.
prompt Type: stringDefault:
Text prompt to send to Kimi K2.6.
max_tokens Type: integerDefault: 2048Range: 1 - 32768
Maximum number of tokens to generate.
image_input Type: arrayDefault:
Optional list of images to include in the prompt. Kimi K2.6 supports vision inputs.
temperature Type: numberDefault: 0.6Range: 0 - 2
Sampling temperature. Higher values make output more random.
system_prompt Type: stringDefault: You are a helpful assistant.
System prompt to guide the model's behavior.
presence_penalty Type: numberDefault: 0Range: -2 - 2
Penalizes new tokens based on whether they appear in the text so far.
reasoning_effort Default: none
How much effort the model spends reasoning before answering. Use 'none' to disable.
frequency_penalty Type: numberDefault: 0Range: -2 - 2
Penalizes new tokens based on their existing frequency in the text so far.
Output Schema

Output

Type: arrayItems Type: string

Version Details
Version ID
a9d565c31037ea4da7300302132940310533fd9e3ace6de17b771206759a6a42
Version Created
April 22, 2026
Run on Replicate →