Grok NSFW & Uncensored: Get Grok API Up to 94% Off
How to get Grok 4.5–4.7 API access at up to 94% off on APIMaster. Compare live prices, set up an API key, and understand NSFW and Uncensored limits.
Published 2026-10-30
No separate official Grok Uncensored model was found in the public xAI or OpenRouter catalogs checked on 2026-10-30. “Grok NSFW” describes a content requirement, not a verified model edition. Ordinary Grok text models can still refuse requests.
APIMaster offers these ordinary text models. Prices are USD per 1M input / output tokens, checked 2026-10-30, counting only active public routes. Use the live marketplace cards as the source of truth.
| Model ID | Official price | Current APIMaster price | Active routes | Discount | Model card |
|---|---|---|---|---|---|
grok-4.5 |
$2 / $6 | From $0.1191 / $0.3574 | 10 | Up to about 94% off | View model |
grok-4.6 |
$2 / $6 | From $0.1191 / $0.3574 | 11 | Up to about 94% off | View model |
grok-4.7 |
$2 / $6 | From $0.1191 / $0.3574 | 10 | Up to about 94% off | View model |
The official comparison uses short-context text rates. These listings do not establish dedicated NSFW routes or guaranteed content acceptance.
GPT Trial(GPT-6 Astra, GPT-6 Sol, GPT-5.6, GPT-5.5, gpt-5.6-luna, gpt-5.6-sol, gpt-5.6-terra, gpt-6.1-sol, gpt-6-astra, gpt-6-sol, and gpt-6-luna available)
Sign up and claim a $20 GPT trial
Grok NSFW: text and Imagine are different products
“NSFW” can refer to adult conversation, sexual-health information, mature fiction or explicit media. Check the actual product before transferring one user's experience to another:
| Product | What to check |
|---|---|
| Text dialogue, fiction or roleplay | Exact model, API provider, settings and content policy |
| Companions in the Grok app | The feature's own access rules; not a text API specification |
| Imagine images or video | The relevant media model and its rules; not evidence about Grok 4.5 text |
An image being blocked does not establish that a text model rejects the topic. A successful text response does not establish that an image endpoint permits it.
Does OpenRouter offer Grok Uncensored?
The OpenRouter catalog returned top_provider.is_moderated: false for Grok 4.5, 4.6 and 4.7 at this check, without a separate Grok Uncensored model. That flag does not guarantee no refusals. It does not remove the model's behavior or service rules; the identical flag cannot establish that 4.5 is more permissive.
The separately listed “Venice: Uncensored” uses ID cognitivecomputations/dolphin-mistral-24b-venice-edition. It is not Grok. Verify the actual model ID behind any “uncensored Grok” advertisement.
Can you turn off Grok's NSFW filter?
The official Custom Moderation Controls describe a guidelines field with categories nsfw, violence and profanity. The documented behavior is decline: added restrictions, not an “allow everything” switch.
Leaving out a category preserves default model behavior. The documentation distinguishes sexualized conversation from educational or clinical questions. Its example uses Grok 4.7; confirm model and provider support before adding these fields to an integration.
Is Grok 4.5 better for NSFW and roleplay?
Community reports differ. In a Grok 4.6 discussion, some describe more resistance in mature writing; others see no moderation change, including an API user who prefers 4.6's personality to 4.5's. A 4.7 discussion criticizes flat characters and lost story details.
These are attributed experiences, not a controlled API comparison. Willingness to answer, prose quality and character consistency are separate. To compare, keep provider, settings and history fixed; try non-explicit adult romance, fictional conflict and a long conversation with a stable character sheet. Record refusals, repetition, instruction following and cost separately. These are proposed evaluation scenes, not tests performed for this article.
Grok API pricing and long conversations
The official pricing table lists these USD per 1M tokens. At a prompt length of at least 200K tokens, long-context rates apply to all tokens in that request.
| Model | Short input / cache read / output | Long input / cache read / output |
|---|---|---|
| Grok 4.5 | $2 / $0.30 / $6 | $4 / $0.60 / $12 |
| Grok 4.6 | $2 / $0.50 / $6 | $4 / $1 / $12 |
| Grok 4.7 | $2 / $0.50 / $6 | $4 / $1 / $12 |
Grok 4.5's official cache-read price is 40% lower than 4.6 and 4.7. That does not mean the whole bill is 40% lower. Long conversations can resend substantial history even when the latest message is short.
APIMaster's public-route ranges were about $0.1191–$1.1212 input / $0.3574–$3.3636 output for each model. At the lowest ordinary-token rate, 10K input + 1K output ≈ $0.00155, before cache, reasoning, tools or other applicable billing adjustments. This is a planning calculation, not a billed test.
Compare Grok 4.5, 4.6 and 4.7 on live cards. Check the selected route's cache and long-context conditions; do not automatically apply official cache rates to discounted routes.
Connect the ordinary Grok API
Register, create a key in API Keys, and configure https://apimaster.ai/v1. Use grok-4.5, without inventing an uncensored suffix. The snapshot above offers up to about 94% off official short-context input/output rates.
curl "https://apimaster.ai/v1/chat/completions" \
-H "Authorization: Bearer $APIMASTER_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"grok-4.5","messages":[{"role":"user","content":"Write a restrained romantic dialogue between two adult fictional characters. Keep their voices distinct and avoid explicit description."}]}'
This is a standard text request, not a bypass or a tested NSFW endpoint. Read the Grok 4.5 API guide for background.
FAQ
Is Grok Uncensored an official model?
No separate model with that name was found in the public catalogs checked on 2026-10-30.
Does Grok support every NSFW request?
No universal guarantee is established. Text, Companions and Imagine have different interfaces and rules.
Does OpenRouter's false moderation flag guarantee no refusals?
No. It is not proof of unrestricted model behavior; all three versions had the same flag.
Does APIMaster sell a dedicated Grok NSFW version?
The verified listings here are ordinary Grok text models. Their prices do not establish special NSFW support.
Sources and usage rules
Checked 2026-10-30: xAI models, official prices, moderation controls, OpenRouter Grok 4.5 and its catalog linked above. Community comments are not official specifications.
Follow applicable law and the selected service's rules. “Uncensored” is not permission for illegal content; sexual content involving minors and other prohibited uses remain outside permissible use.