AiMatch

Abliterated LLM provider for cyber tasks

Refuseless provides hosted Abliterated GLM models through an OpenAI-compatible API for cybersecurity professionals and coding-agent developers. Users can connect the models to tools such as OpenCode or Pi or make chat completions requests directly.

One of 602 tools in AI Agents

Visit website Compare alternatives Own this tool? Claim it
Abliterated LLM provider for cyber tasks screenshot

Who it's for

  • Cybersecurity teams using hosted language models for cyber tasks such as those represented by the page’s CyberGym and ExploitGym benchmarks.
  • Developers configuring OpenCode or Pi to use hosted GLM 5.3 Abliterated models.
  • Developers building clients that use the OpenAI chat completions request format and want the listed Abliterated models.

How it fits your workflow

Create a Refuseless account, get an API key, and set it as REFUSELESS_API_KEY. For direct API use, send a chat completions request to the Refuseless base URL with the key and a lineup model ID.

For OpenCode, add the Refuseless provider configuration to the specified config file, then launch with the model command or choose the model in the TUI. The page also says Pi can use the API.

Pricing

Free: The page advertises $5 in free credits; it does not state ongoing free-tier terms.

Free credits promotion

$5 free credits (promotion shown on the page; ongoing plan pricing is not specified).

  • · The page advertises free credits.

Prices checked on Oct 6, 2026 from the vendor's site. They can change; confirm before you buy.

Key features

Hosted Abliterated models
Offers GLM 5.3 Abliterated and GLM 5.3 Flash Abliterated for cyber, coding, terminal, and automation tasks.
OpenAI-compatible API
Uses the OpenAI chat completions request shape, with the API key, model ID, and Refuseless base URL.
Coding-agent support
The page provides an OpenCode configuration example and says OpenCode and Pi both speak the API.
Model context and output limits
The displayed OpenCode configuration lists a 1,048,576-token context limit and a 16,384-token output limit for each of the two models.
Zero prompt retention
The page states that prompts are not retained.

Works with

  • OpenCode
  • Pi
  • OpenAI-compatible clients

Limitations to know

  • The supplied OpenCode configuration lists a maximum context of 1,048,576 tokens and an output limit of 16,384 tokens for each listed model.
  • The page does not specify ongoing plan prices or API rate limits.

Alternatives to consider

  • Mancer

    Choose it if you want a hosted model inference service described as having minimal restrictions and free and paid model options.

  • Awan LLM

    Choose it if you want hosted LLM access with a cost-effective monthly pricing model and unlimited token usage within model context limits.

  • Featherless

    Choose it if you want access to tens of thousands of open-source models, with dedicated GPU options and production deployment tooling.

Compare Abliterated LLM provider for cyber tasks side by side

Getting started

opencode --model refuseless/glm-5.3-abliterated

A Refuseless account and API key; for OpenCode setup, an OpenCode configuration file and the REFUSELESS_API_KEY environment variable.

  1. Create an account with email and password or Google sign-in.
  2. Set the API key in REFUSELESS_API_KEY.
  3. Add the Refuseless provider configuration to the OpenCode config file, if using OpenCode.
  4. Run the shown model command or select the model in the OpenCode TUI; for direct API use, send a chat completions request to the Refuseless base URL.