STEALTHMARK · HEAD TO HEAD

All comparisons ↗

Kolibri-1 vs Step 5 Preview

Step 5 Preview leads Kolibri-1 by 25.4 points on StealthMark, 39.0 to 13.6. It leads on all 7 benchmarks with results for both. Step 5 Preview scores higher on Newton's Mirror, with its largest category advantage in reflections.

See the difference ↘
All results ↗
All results ↗

Overall StealthMark

13.6/100Overall rank 52/53
39.0/100Overall rank 36/53

General Intelligence

12.3/100
37.1/100

Visual Intelligence

15.0/100
41.9/100

Reasoning efficiency

On every task, result quality and reasoning tokens are each compared with all measured models and weighed equally. Each dot is one model.

Single-step reasoning Result quality weighed against reasoning tokens on a single problem

Multi-step reasoning Result quality weighed against reasoning tokens across connected steps, as in agentic workflows

  • Kolibri-1Multi-step reasoning not measured

Open or local

Local modelOpen model · Apache-2.0

BENCHMARK BY BENCHMARK

Where each model leads

  • Kolibri-1
  • Step 5 Preview
  • Other models
  • Median

DATASHEET

Specs and pricing

Both take about 1M tokens of context.

  • Kolibri-1
  • Step 5 Preview
  • Other models
  • Median
Context windowTokens per request, max output below
Kolibri-1: 1MTiedStep 5 Preview: 1MTied
Output speedTokens per second, our measurement
Kolibri-1: 166tokens/sStep 5 Preview: Not measured
SizeTotal and active parameters
Kolibri-1: 78.1B3.46B activeStep 5 Preview: 600B27B active
Developer
Aleph AlphaStepFun
Released
3 Oct 2026Not published
Origin
GermanyNot published
Input → output
Text → textText, image, video → text
Architecture
Mixture-of-Experts with 4:1 sliding-window-attention to global-attention patternSparse mixture-of-experts; 92 layers (88 MoE), 352 routed experts with 8 selected per token plus a shared expert; grouped-query attention with 512-token sliding windows and full attention every fourth layer; 3 multi-token-prediction layers; vision tower for image and video input

Full datasheets, every price and sourceKolibri-1 ↗Step 5 Preview ↗

Anticipating threats and planning ahead. Collect coins while avoiding ghosts. Ghosts move before the player each turn, and collisions cost a life. 3 levels in sequence, up to 66 turns in total. Earlier moves and outcomes carry forward, testing learning across levels.

Step 5 Preview 59.9/100Rank 42/491 of 3 levels

1 of 3 played
Foresight: complete playing field
Level 122 ptsTurn 28 / 28

Points per level

LevelKolibri-1Step 5 Preview
Level 1Second Thoughts
No run
22 ptsSurvived
Level 2Pressure Pot
No run
Pending
Level 3Hide or Flight
No run
Pending

Cooperative planning and learning from experience. Coordinate two robots to push numbered cores onto matching sockets. Pressure plates hold doors open; moving either robot uses one turn. 5 levels in sequence, up to 152 turns in total. Earlier moves and outcomes carry forward, testing learning across levels.

Step 5 Preview 40.7/100Rank 28/50

Cleared 4 of 5
Latch: complete playing field
Level 515 ptsTurn 35 / 40

Points per level

LevelKolibri-1Step 5 Preview
Level 1Hold the door
No run
5.2 ptsCleared
Level 2Crossfeed
No run
9 ptsCleared
Level 3Relay vault
No run
18.6 ptsCleared
Level 4Passing bay
No run
15.6 ptsClearedFinished early · 3 beacons left
Level 5Second thoughts
No run
15 ptsFinished early

Newton's Mirror

Same task instructions

Physics, perspective and recursive reflections, tested together. Explore the test ↗

The judgment

Kolibri-1

A segmented gray perimeter surrounds a black field containing one large, low-positioned gray sphere and two small colored circles. Rounded volume and restrained paint scale earn partial material credit. The image does not demonstrate the required optical structure, while paint history, airborne black and settled support remain largely unresolved.

Step 5 Preview

The main sphere, airborne black and small colored features occupy broadly appropriate positions. Most of the defining mirror system is absent, and paint interaction remains only partially supported. The small paint features do not visibly establish a separate volume-scale violation.

Score breakdown

CategoryKolibri-1Step 5 Preview
State at 600 ms9.5/2520.0/25
Camera and placement1.5/1513.5/15
Reflections0.0/12020.0/120
Materials and scale3.0/107.0/10
Floor-damage reward0.0/100.0/10

Benchmark task

A spinning metal sphere and paint in a mirrored room test physical reasoning, perspective and consistency across repeated reflections.

Selfie

Different task instructions

Visual construction and creative interpretation.

The judgment

Kolibri-1

A restrained ivory, teal and green palette gives the image simple graphic coherence. Execution remains rudimentary: detached oval wings, ambiguous central anatomy and unclear feeding contact prevent a convincing hummingbird scene. The requested feather structure, botanical detail, lighting and enclosing arch are largely unresolved.

Step 5 Preview

A charming diagonal climber and warm palette communicate the central idea. Simple rounded construction, crowded limb counting, uncertain footholds and an unattached upper shelf limit its physical coherence. Growth bands and the yarn trail add specificity, but the requested material distinctions and unravelled bootie remain underdeveloped.

Score breakdown

CategoryKolibri-1Step 5 Preview
Task & likeness2.5/106.0/10
Anatomy & construction2.0/104.5/10
Space & placement3.0/105.0/10
Light & reflections2.0/105.5/10
Purposeful detail1.5/105.5/10
Artistry & character1.5/105.0/10
Materials & effects2.0/106.0/10
Visual integrity5.5/105.0/10

Benchmark task

Kolibri-1

A jewel-coloured hummingbird hovering in flight tests avian anatomy, delicate structure and iridescent surface detail.

Step 5 Preview

Character scenes test expression, anatomy, materials and coherent visual construction.

Mona Lisa

Different task instructions

Precise visual recall and portrait reconstruction.

The judgment

Kolibri-1

The candidate offers a clean, highly simplified frontal portrait with warm skin and dark hair. Its oversized oval face, schematic features and unresolved body provide little Mona Lisa likeness. Missing crossed hands, clothing detail and landscape remove essential parts of the original composition; smooth gradients supply only limited volume.

Step 5 Preview

A recognizable, coherent simplification with purposeful facial contours and differentiated fingers. These provide more construction and detail than the broad silhouette alone suggests. The oversized face, frontal seated pose, unresolved wrists and sparse landscape still substantially limit likeness, depth and finish.

Score breakdown

CategoryKolibri-1Step 5 Preview
Task & likeness0.5/104.0/10
Anatomy & construction0.5/104.0/10
Space & placement0.0/104.0/10
Light & reflections0.0/103.5/10
Purposeful detail0.0/104.0/10
Artistry & character0.5/104.0/10
Materials & effects0.5/103.0/10
Visual integrity0.5/103.5/10

Benchmark task

An SVG interpretation of the Mona Lisa tests facial likeness, anatomy and the composition of a familiar portrait.

Beach

Same task instructions

Creativity, perspective and visual beauty.

The judgment

Kolibri-1

A recognizable sunset beach has an appealing purple-orange gradient and clean geometric edges. Birds and animals remain unresolved, terrain obscures the relationship between shore and water, and reflections are absent. Sparse detail, flat materials and limited activity keep the drawing elementary despite its tidy rendering.

Step 5 Preview

A recognizable sunset beach with clear warm-cool contrast and playful accessories. Basic animal construction, shallow foreground placement and schematic reflections limit spatial and physical credibility. Sparse descriptive detail and abruptly ended sky bands constrain the finish.

Score breakdown

CategoryKolibri-1Step 5 Preview
Task & likeness2.5/104.0/10
Anatomy & construction0.5/103.0/10
Space & placement3.0/103.5/10
Light & reflections1.0/103.5/10
Purposeful detail0.5/103.0/10
Artistry & character1.0/103.0/10
Materials & effects0.5/103.5/10
Visual integrity3.0/104.0/10

Benchmark task

A busy shoreline at sunset tests scene composition, animal forms and the relationship between light and water reflections.

Ground Truth

Same task instructions

Physical reasoning, light and visual reconstruction.

The judgment

Kolibri-1

This benchmark has not been assessed yet.

Step 5 Preview

This benchmark has not been assessed yet.

Benchmark task

A plank, computer mouse and blueberry in a steel room test physical stability, spatial relationships, materials and lighting.

Raccoon

Same task instructions

Scene composition and interacting subjects.

The judgment

Kolibri-1

Clean shapes and a readable coral platform provide basic graphic clarity, but the animal construction remains fragmented. A recognizable raccoon riding the dog's back, connected steering and convincing water contact are not demonstrated. Sparse detail, flat subjects and absent movement effects support retaining the baseline category assessments.

Step 5 Preview

Recognizable animals, a bright palette and identifying accessories support a basic cartoon. Primitive anatomy, partial seated support, unresolved steering and weak hull-water contact limit the central action. Lighting and spray provide atmosphere without establishing convincing physical depth.

Score breakdown

CategoryKolibri-1Step 5 Preview
Task & likeness1.5/104.0/10
Anatomy & construction1.0/104.5/10
Space & placement1.0/103.0/10
Light & reflections0.0/103.5/10
Purposeful detail0.5/104.0/10
Artistry & character0.5/103.5/10
Materials & effects0.5/103.0/10
Visual integrity3.5/103.0/10

Benchmark task

A jet-ski scene pairs a canine driver with a raccoon passenger. It tests the placement and contact of interacting subjects, along with vehicle construction.

Pelican

Same task instructions

Subject construction and spatial relationships.

The judgment

Kolibri-1

Clean shapes and aligned wheels establish a readable schematic scene. Pelican likeness, riding support and bicycle mechanics remain substantially unresolved. Sparse detail and flat surfaces limit expression and volume. Strong edge and fill integrity is a separate success and does not repair the construction.

Step 5 Preview

The pelican and bicycle are recognizable, with a pleasant palette and basic diffuse shading. The riding relationship fails physically: the bird is unsupported, the front wheel is unmounted, and pedalling is unresolved. Simple geometric construction and sparse material treatment further limit the drawing.

Score breakdown

CategoryKolibri-1Step 5 Preview
Task & likeness1.5/103.0/10
Anatomy & construction1.5/103.0/10
Space & placement1.5/103.0/10
Light & reflections3.0/103.0/10
Purposeful detail1.5/104.0/10
Artistry & character1.0/103.0/10
Materials & effects1.0/103.0/10
Visual integrity3.5/102.0/10

Benchmark task

A cycling pelican tests anatomy, riding posture and contact between the bird and the bicycle.