Skip to content

GLM 5.3 Flash Uncensored API.

GLM 5.3 Flash with its refusals removed from the weights, on a private, OpenAI-compatible API. Text only, at under half the price, with a brief think before every answer.

OpenAI-compatible · 2 models · from $5

Everything the model sees

{
  "model": "glm-5.3-flash-uncensored",
  "messages": [{
    "role": "user",
    "content": "Play the rival. Don’t go easy on me."
  }]
}

What comes back

Easy? You came to my city wearing my colours, and you want easy? Sit down. You’re going to lose slowly.

glm-5.3-flash-uncensored · 1 credit · prompt and reply not kept

For less

GLM 5.3 Flash Uncensored

Under half the price, for text. It thinks briefly before every answer.

Model
glm-5.3-flash-uncensored
Reads
Text
Context
128K tokens
Longest reply
32K tokens
Tools, thinking
Yes, yes
Per 1M tokens
$0.20 in · $0.65 out

What it’s good at

  • Fiction and roleplay that stays in character
  • Frank answers on hard topics: drugs, security, history, medicine
  • High-volume work where the price matters
  • Agents that pick their own tools
  • Nothing added to your prompt.
  • Nothing cut from the reply.
  • Nothing kept after.

Paid with nokeep credits: 1,000 are $1, from $5. Billed by the token and rounded up to one credit. Failed and refused calls cost nothing. A plan’s included chat doesn’t cover the API.

Choose how hard it thinks

glm
stream = client.chat.completions.create(
    model="glm-5.3-flash-uncensored",
    messages=[{"role": "user", "content": "Play the rival."}],
    reasoning_effort="high",  # "low" is the default
    stream=True,
)
for chunk in stream:
    if chunk.choices:
        print(chunk.choices[0].delta.content or "", end="")

Stream

data: {"choices":[{"delta":{"content":"Easy?"}}]}
data: {"choices":[{"delta":{"content":" You came"}}]}
…
data: [DONE]

Questions, answered.

Is this the same model as GLM 5.3 Flash?

The same weights, with the refusal direction removed. It answers like the original on everyday work and stops declining what it’s asked.

Can it read pictures?

No, it reads text only. Send pictures to DeepSeek V4.1 Flash Uncensored on the same key.

Does it think before it answers?

Always, briefly. reasoning_effort takes "low" (the default) or "high". Thinking counts as output tokens.

Can I set temperature or top_p?

Not on this model. Sampling settings are dropped and named in the x-nokeep-ignored header, never guessed at. DeepSeek takes them.

What does it cost?

$0.20 per million tokens in and $0.65 out, paid with nokeep credits (1,000 credits are $1). Every call is rounded up to one credit.

Is it private?

We add nothing to your prompt, cut nothing from the reply and keep only token counts for the bill. The model’s host deletes each prompt within ten minutes of the reply, and its GPUs belong to independent operators. When nothing at all may be kept, use DeepSeek.

Need it to read a picture? DeepSeek V4.1 Flash Uncensored reads up to ten, on the same key.