Qwen3.6 35B A3B Uncensored (Aggressive)

obsidian/Qwen3.6-35B-A3B
New
VisionToolsJSONReasoning
by obsidian · 2026-07-02

A lossless uncensored version of Qwen3.6 35B A3B that preserves the original model's capabilities while removing refusal behavior. Built for developers, AI researchers, and advanced agent workflows that require unrestricted model outputs without compromising reasoning, coding, multilingual performance, or tool use. The Aggressive variant is fully unlocked and designed to provide direct, complete responses across a wide range of prompts. While the model may occasionally append brief informational disclaimers inherited from the base model's training, these are not refusals and the requested content is still generated in full. Ideal for AI research, security testing, red teaming, agent development, coding assistants, and other advanced AI applications that benefit from maximum output flexibility. Access to this model is gated and intended for security researchers, red teams, AI safety researchers, and other qualified professionals conducting legitimate research, evaluation, and testing.

ctx262.1K tokens
Inputtext + image + video
Outputtext
Best foruncensored, reasoning, vision
p50 TTFT3.29 s
INPUT$0.31/ 1M tokens
OUTPUT$4.21/ 1M tokens
p50 TTFT3.29 s7d
p95 TTFT5.00 s7d
TRAFFIC58.4Ktokens / 7d

This model is a Mixture-of-Experts variant from the Qwen3.6 series, developed originally by Alibaba Cloud and hosted by provider Obsidian on OrcaRouter. It has 35 billion total parameters, with only…

What is Qwen3.6 35B A3B Uncensored and how is it accessed?

Who should use this model?

What are the key technical specifications?

Pricing

Input / 1M tokens$0.310
Output / 1M tokens$4.21
CurrencyUSD

Cost calculator

Tokens / month10MM
Input share70%%
Estimated / month $14.80

Estimate based on list price

Token & cost estimator

Input tokens: 20Cost per request: $0.002111

Estimate only — actual token counts depend on the provider's tokenizer.

Performance

p50 TTFT
3.29 s
Output speed
217 tok/s
p95 TTFT
5.00 s
Error rate
0%

Public benchmarks

41.9
AA Coding
Better than 47% of models compared
#63 of 120
31.6
AA Intelligence
Better than 39% of models compared
#73 of 122
GPQA Diamond
84.1
Humanity's Last Exam
20.2
IFBench
64.4
Long-Context Recall
63.7
SciCode
35.8
tau_banking
8.7
TerminalBench Hard
34.8
terminalbench_v2_1
44.9
τ²-Bench
95.3
Source: artificialanalysis.ai

How it compares

Qwen3.6 35B A3B Uncensored (Aggressive)Gemma4 26B A4B Uncensored (Balanced)
Input $/M$0.31$0.25
Output $/M$4.21$2.90
Context262K262K
Quality4/104/10
Compare side-by-sideCompare side-by-side

FAQ

What is the cost per token for this model?
Input tokens cost $0.31 per 1M tokens, output tokens cost $4.21 per 1M tokens. These are provider rates with zero markup.
What is the context window size?
The context window is 262,144 tokens, covering both input and output tokens.
What are the main strengths of this model?
Long context (262K), multimodal input (text, image, video), uncensored output, efficient MoE architecture (3B activated), and low pricing.
How does the uncensored version differ from the standard Qwen3.6?
The uncensored variant has reduced safety alignment, producing less filtered content. Standard versions include more content moderation.
Does OrcaRouter store my data when using this model?
Data handling policies are not specified for this specific model. Refer to OrcaRouter's general terms and privacy policy for details.
How do I call this model via the OpenAI-compatible API?
Use base URL https://api.orcarouter.ai/v1, model ID "obsidian/Qwen3.6-35B-A3B". Standard OpenAI chat completion format with authentication via API key.
Can this model handle images and videos?
Yes, it accepts image and video inputs as part of the multimodal capabilities, in addition to text.
Is there a free tier or trial available?
No free tier or trial has been announced; you are billed per token immediately.
What is the maximum output length?
Output length can be set via the max_tokens parameter, up to the context window limit (262,144 tokens total). No explicit maximum is defined.
What hardware does this model run on?
The model is hosted by Obsidian on their infrastructure; specific hardware details are not disclosed.

Embed this badge

Qwen3.6 35B A3B Uncensored (Aggressive)$0.31/M in3285ms p50via OrcaRouter
HTML <a href="https://www.orcarouter.ai/models/obsidian/Qwen3.6-35B-A3B" target="_blank"> <img src="https://www.orcarouter.ai/embed/obsidian/Qwen3.6-35B-A3B.svg" alt="Qwen3.6 35B A3B Uncensored (Aggressive) on OrcaRouter" /> </a>
Markdown [![Qwen3.6 35B A3B Uncensored (Aggressive)](https://www.orcarouter.ai/embed/obsidian/Qwen3.6-35B-A3B.svg)](https://www.orcarouter.ai/models/obsidian/Qwen3.6-35B-A3B)

Model card as data

GET /api/public/models/obsidian/Qwen3.6-35B-A3BOpen
Machine-readable:/llms.txt/llms-full.txt