GLM-5.3 Flash Abliterated API: $0.50/$1.60 Setup

GLM-5.3 Flash Abliterated API pricing on RouterPlex is $0.50 per 1M input tokens and $1.60 per 1M output tokens. Cached input is $0.30, and the model id is glm-5.3-flash-abliterated. Checked 23 October 2026. Open GLM-5.3 Flash Abliterated for the live rate.

This is GLM-5.3 Flash with refusal behaviour removed from the weights. It is the low-cost route for security work the stock model declines, such as large red-team prompt runs. RouterPlex sets this rate, and it is not the Z.ai list price.

RouterPlex publishes this page and sells the prepaid wallet described below.

Create a free RouterPlex account and add a $5 top-up. Cap the key, then send one short request before you move a real agent onto this id.

What it is for #

Stock models often refuse legitimate security work. A refusal-removed model answers it. Teams red-team their own LLM apps and test guardrails with real attack prompts. They write penetration-test reports for systems they are authorized to test. They ask how a disclosed CVE works, then draft detection rules and patches for it. Incident responders use it to read suspicious scripts and logs. Others build evaluation sets on sensitive topics without the model stopping halfway.

You are still bound by the terms. Authorized work is your responsibility, and the model can be wrong, so check what it gives you.

GLM-5.3 Flash Abliterated API pricing #

USD per 1M tokens.

Z.ai ran a stock-flash promo at $0.075 and $0.25 through 9 September 2026. RouterPlex bills the standard stock list rate of $0.15 and $0.50 on glm-5.3-flash, and the promo price is not the abliterated price either. The abliterated flash rate is the RouterPlex figure in the first column.

A cold call of 40,000 input tokens and 2,000 output tokens costs $0.0232 here, and the same mix is $0.0070 on stock flash and $0.172 on glm-5.3-abliterated. With 30,000 of the 40,000 tokens cached, the abliterated flash bill is $0.0172, and that is $0.005 fresh, $0.009 cached, and $0.0032 output.

A $5 balance covers 10 million fresh input tokens at $0.50, or 3.125 million output tokens at $1.60. Real calls mix both, and reasoning text adds output tokens. The model page lists reasoning effort as low, high, or max. This article did not measure how many extra tokens each effort adds.

Why run it on RouterPlex #

One prepaid key calls this route next to Claude, GPT, DeepSeek, and stock GLM, so a red-team test suite can compare refusals across models without a second account. Top-ups have a $0 fee, and the minimum is $5. The balance stops at $0, and each key can carry a hard spend cap, which keeps a large prompt run from overspending.

The full abliterated GLM, at $4 and $6, is the stronger sibling. Its page is GLM-5.3 Abliterated. Stock flash setup and the vendor list rate are in GLM-5.3 API pricing.

Specs #

Context is 1,048,576 tokens. Max output is 131,072. The catalog marks tool calls, image input, and reasoning on. The public page slug is /models/glm-5-3-flash-abliterated. The dotted URL /models/glm-5.3-flash-abliterated returns 404. The JSON model id keeps the dots.

Setup #

curl https://api.routerplex.com/v1/chat/completions \
  -H "Authorization: Bearer $ROUTERPLEX_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "glm-5.3-flash-abliterated",
    "messages": [
      {"role": "user", "content": "Reply with exactly: RouterPlex connected."}
    ]
  }'

Claude Code and other Anthropic clients use the gateway root, not the /v1 chat path.

export ANTHROPIC_BASE_URL=https://api.routerplex.com
export ANTHROPIC_AUTH_TOKEN=$ROUTERPLEX_API_KEY
export ANTHROPIC_MODEL=glm-5.3-flash-abliterated

The OpenAI-compatible API guide covers the SDK base URL https://api.routerplex.com/v1. Claude Code Router covers agent routing. Keep the key in the environment. Do not commit it.

Errors #

The $1 trial key is not free access to this id, and there is no free credit on abliterated routes. HTTP 402 with type paid_model_required means the key is still a trial key. Add a paid top-up of at least $5, and the top-up fee is $0.

A hard spend cap stops the key even when the account balance remains. Raise the cap or split the job, and the account itself does not go negative. A negative-balance 402 on another gateway is a different bug, covered in Fix an OpenRouter 402.

Model not found means the id string is wrong. Send glm-5.3-flash-abliterated, not the page slug and not glm-5.3-flash.

Terms #

The model will answer some requests that stock GLM-5.3 Flash declines, and you are responsible for that use under the terms. Illegal content, malware, weapons development, and child sexual abuse material are barred.

Next step #

Create a free RouterPlex account, add a $5 top-up, and open GLM-5.3 Flash Abliterated before you set a budget.