stealthmodels.

MODEL FIELD GUIDE

Qwen3.6 35B A3B: Newton's Mirror

A multimodal 35B-parameter MoE that activates 3B parameters per token. It combines open Apache-2.0 weights with an emphasis on coding agents.

Drawn by this model01 / 01
Newton's Mirror benchmark result generated by this model
SVG BENCHMARKNewton's MirrorView result

A good fit for

Local or hosted coding assistants that need image understanding and modest active compute. Plan memory for the full checkpoint.

Context window
1,010,000tokens[1]
Active parameters
3Bparameters[1]
License
Apache-2.0[1]
Model factsArchitecture, context & capabilities
Developer
Qwen[1] [2]
Origin
China (Alibaba Group/Qwen organization context)[3]
Family
Qwen 3.6[1] [2]
Context window
1,010,000 tokens[1]
Maximum output
65,536 tokens[2]
Architecture
MoE causal language model with vision encoder; Gated DeltaNet + gated attention; 256 experts, 8 routed + 1 shared active[1]
Total parameters
35B[1]
Active parameters per token
3B[1]
Input
text, image, video[1] [2]
Output
text[1] [2]
API access & pricingOpenRouter · $0.15 input / $1 output · per 1M tokens

Alibaba Cloud Model Studio; input 0.248 USD/1M; output 1.485/1M; Representative documented rate; region-specific prices exist.

[2]
  • OpenRouterObserved 2026-10-06
    Input $0.15Output $1Cached read $0.05

    per 1M tokens

    qwen/qwen3.6-35b-a3b[4]
  • Alibaba Cloud Model StudioObserved 2026-10-06
    Input $0.248Output $1.485

    per 1M tokens

    Available

    qwen3.6-35b-a3b[2]
Run it locallyWeights, memory & deployment

high-memory single GPU / small multi-GPU

[1] [6]
Measurements & sources6 primary references · Speed & serving conditions

StealthMark score

StealthMark measures fluid and visual intelligence of AI models.

General and Visual

General Intelligence

Perspective, anatomy, correct shadows, purposeful detail and physical consistency.

Visual Intelligence

Likeness, anatomy, detail, light, materials, composition and artistic expression.

Judging

Frontier Models blindly judge the AI generation unbiased through an intricate process. Our researchers control for mistakes.

Overall score

Overall combines weighted task scores on a scale from 0 to 100. Repeating a task does not increase its weight. Estimated means the total includes an estimate for one missing scene.

Benchmaxxing

Benchmaxx flags models that score unusually well on the familiar Pelican task compared with their other benchmark results. The percentage measures that imbalance, not the probability of training contamination. Flagged Pelican scores are excluded from Overall. Pelican never counts toward General or Visual.

We keep some prompts private to discourage test-specific optimization.

Scoring categories

Prompt & likeness

Subjects, actions and likeness.

Anatomy & construction

Coherent bodies, joints and machinery.

Space & placement

Perspective, scale and contact.

Light & reflections

Lighting, shadows and reflections.

Purposeful detail

Clear, purposeful small features.

Artistry & character

Composition, expression and drawing skill.

Materials & effects

Surfaces, texture and movement.

Visual integrity

Clean shapes, layers and edges.

BENCHMARK RESULTS

Compare results

This link opens the same results in the same order.

StealthMark benchmark results

Reset benchmark preferences?

Return to the default view, clear comparison selections and restore the advertisement.