stealthmodels.

MODEL FIELD GUIDE

Minimax M3: Newton's Mirror

Open-weight multimodal MoE with 428B total parameters, 23B active and a million-token context, under MiniMax's community license.

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

A good fit for

Long-context multimodal agents through a low-cost API, or self-hosting on large servers under the model's license terms.

Context window
1,000,000tokens[1] [2]
Active parameters
23Bparameters[2]
License
MiniMax Community License[3]
Model factsArchitecture, context & capabilities
First available
2026-06-01[1]
Developer
MiniMax[1] [2] [3]
Origin
Singapore (MiniMax API service entity Nanonoble Pte. Ltd.)[4]
Family
MiniMax M3[1] [2] [3]
Context window
1,000,000 tokens[1] [2]
Architecture
Mixture-of-Experts with MiniMax Sparse Attention and native multimodal encoder[2]
Total parameters
428B[2]
Active parameters per token
23B[2]
Input
text, image, video[1] [2]
Output
text[1] [2]
API access & pricingOpenRouter · $0.3 input / $1.2 output · per 1M tokens

MiniMax API; input 0.3 USD/1M; cached read 0.06/1M; output 1.2/1M; Displayed permanent 50% discount: <=512K $0.30/$1.20; 512K-1M $0.60/$2.40. List rates are twice those values.

[6]
  • OpenRouterObserved 2026-10-06
    Input $0.3Output $1.2Cached read $0.06

    per 1M tokens

    minimax/minimax-m3[7]
  • MiniMax APIObserved 2026-10-06
    Input $0.3Output $1.2Cached read $0.06

    per 1M tokens

    Available

    MiniMax-M3[6]
Run it locallyWeights, memory & deployment

multi-GPU server

[2] [5]
  • Unsloth UD-IQ1_M128 GB storage · ~1-bit[5]
  • Unsloth BF16852 GB storage · BF16[5]
Measurements & sources8 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.