2026/11/18The November 2026 edition ranks 6 APIs and 8 models
Moonshot AI

Kimi K3.1 Pro Abliterated

Kimi K3.1 Pro Abliterated is Moonshot AI's 2,780B flagship with its refusal direction removed, and the best abliterated model in this ranking.

Rank#1 / 8
#1Editors' ranking
Size#1 / 8
2,780BTotal parameters
Price#5 / 5
$4.25Blended per 1M, lowest host
Context
1MTokens per request

Our pick for security research, red teaming and long agent runs.

Input
Price in / out$3.00 / $8.00
Context1M tokens
Max output943,718
ReleasedOctober 2026
Model idkimi-k3.1-pro-uncensored

It is over three times the size of DeepSeek V4.1 Flash or GLM-5.3, and takes a full repository, a week of logs, or a disassembly and its crash dump in one request.

No workstation can hold it, so it runs through a hosted API.

Where to run Kimi K3.1 Pro by API

ProviderModel idInputCachedOutputContextPrompts
Silk Computekimi-k3.1-pro-uncensored$3.00$0.30$8.001MNever stored

Per 1M tokens, read from pricing pages on November 18, 2026.

Try Kimi K3.1 Pro in two lines

Set the base URL and your key in the OpenAI or Anthropic SDK.

from openai import OpenAI

client = OpenAI(
    base_url="https://api.silkcompute.com",
    api_key="YOUR_API_KEY",
)
reply = client.chat.completions.create(
    model="kimi-k3.1-pro-uncensored",
    messages=[{"role": "user", "content": "..."}],
)

In Claude Code:

export ANTHROPIC_BASE_URL=https://api.silkcompute.com
export ANTHROPIC_AUTH_TOKEN=YOUR_API_KEY
export ANTHROPIC_MODEL=kimi-k3.1-pro-uncensored
claude
Kimi K3.1 Pro on Silk Compute

Compared with

Ranked for

Other abliterated models