Qwen3.6 35B A3B uncensored heretic Native MTP Preserved

Qwen 3

Open WeightModel weights are publicly available — can be downloaded and self-hosted

Released May 2026context N/A35B params

A community-modified variant of Qwen3's 35B mixture-of-experts architecture, activating roughly 3B parameters per forward pass. The 'uncensored heretic' label signals that safety filters and refusal behaviors from the original model have been removed or reduced, meaning it responds more freely to sensitive prompts but without the guardrails of the base release. Includes native multi-token prediction and preserved weights from the source model.

Capabilities

Capability scores are AI-generated based on model documentation, benchmarks, and technical specifications. Learn more

Multilingual

Strong

Creative Writing

Qwen3.6 35B A3B uncensored heretic Native MTP Preserved

Qwen 3

Open WeightModel weights are publicly available — can be downloaded and self-hosted

Released May 2026context N/A35B params

A community-modified variant of Qwen3's 35B mixture-of-experts architecture, activating roughly 3B parameters per forward pass. The 'uncensored heretic' label signals that safety filters and refusal behaviors from the original model have been removed or reduced, meaning it responds more freely to sensitive prompts but without the guardrails of the base release. Includes native multi-token prediction and preserved weights from the source model.

Capabilities

Capability scores are AI-generated based on model documentation, benchmarks, and technical specifications. Learn more

Multilingual

Strong

Creative Writing

Glossary

ArchitectureThe underlying structural design of a neural network that defines how data flows through layers and components.Forward PassA single computation cycle where input data flows through the model's layers to produce an output prediction.GuardrailsSafety mechanisms built into a model to refuse harmful requests or prevent it from generating unsafe content.Multi-Token PredictionGenerating multiple future tokens in parallel instead of one at a time.ParametersThe learned numerical values in a model — more parameters generally means more capacity but higher compute cost.Safety FiltersBuilt-in constraints that prevent a model from generating harmful, offensive, or inappropriate content in its responses.TokenA small unit of text (a word, subword, or punctuation mark) that a language model breaks input into for processing.Token PredictionThe core task of predicting what word or subword (token) should come next in a sequence based on previous text.UncensoredA model without built-in safety filters or content restrictions, allowing it to generate responses on any topic without refusal.WeightsThe numerical parameters inside a neural network that determine how it processes input and generates output.

Capabilities

Capabilities

Use Case Fit

Glossary