stealthmodels.

MODEL FIELD GUIDE

Kimi K3: Pelican

Open-weight multimodal reasoning at a 2.8T-parameter scale, with a million-token context and 104B parameters active per token.

Drawn by this model01 / 01
Pelican benchmark result generated by this model
SVG BENCHMARKPelicanView result

A good fit for

Complex research and agent workflows combining long documents and images. Hosted access avoids the large memory footprint of self-hosting.

Context window
1,048,576tokens[1]
Active parameters
104Bparameters[1]
License
Kimi K3 License[2]
Model factsArchitecture, context & capabilities
Developer
Moonshot AI[1] [2]
Origin
China[3]
Family
Kimi K3[1] [2]
Context window
1,048,576 tokens[1]
Architecture
Mixture-of-Experts with KDA + Gated MLA attention, 896 experts and native vision encoder[1]
Total parameters
2.8T[1]
Active parameters per token
104B[1]
Input
text, image[1]
Output
text[1]
API access & pricingOpenRouter · $0.69 input / $15 output · per 1M tokens
  • OpenRouterObserved 2026-10-06
    Input $0.69Output $15Cached read $0.23

    per 1M tokens

    moonshotai/kimi-k3[5]
Run it locallyWeights, memory & deployment

rack-scale

[1] [4]
  • Unsloth UD-IQ1_M GGUF649 GB storage · ~1-bit[4]
  • Unsloth UD-Q4_K_XL GGUF1,510 GB storage · ~4-bit[4]
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.