STEALTHMARK · HEAD TO HEAD

All comparisons ↗

DeepSeek-V4-Pro-0813 vs Step 5 Preview

DeepSeek-V4-Pro-0813 leads Step 5 Preview by 2.8 points on StealthMark, 41.8 to 39.0. It leads on 6 of 9 benchmarks, including both games, and Step 5 Preview on 3. DeepSeek-V4-Pro-0813 scores higher on Beach, with its largest category advantage in prompt & likeness.

See the difference ↘
All results ↗
All results ↗

Overall StealthMark

41.8/100Overall rank 33/53
39.0/100Overall rank 36/53

General Intelligence

41.6/100
37.1/100

Visual Intelligence

42.3/100
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

Open or local

Open modelMIT

BENCHMARK BY BENCHMARK

Where each model leads

  • DeepSeek-V4-Pro-0813
  • Step 5 Preview
  • Other models
  • Median

DATASHEET

Specs and pricing

Both take about 1M tokens of context.

  • DeepSeek-V4-Pro-0813
  • Step 5 Preview
  • Other models
  • Median
PricePer 1M output tokens, input below
DeepSeek-V4-Pro-0813: $3.82$1.27 inputStep 5 Preview: No list price
Context windowTokens per request, max output below
DeepSeek-V4-Pro-0813: 1MTied384K max outputStep 5 Preview: 1MTied
Output speedTokens per second, our measurement
DeepSeek-V4-Pro-0813: 137tokens/sStep 5 Preview: Not measured
SizeTotal and active parameters
DeepSeek-V4-Pro-0813: 1.6T49B activeStep 5 Preview: 600B27B active
Developer
DeepSeek-AIStepFun
Released
13 Aug 2026Not published
Origin
ChinaNot published
Input → output
Text → textText, image, video → text
Architecture
Mixture-of-Experts with DeepSeek V4 architecture and DSpark speculative-decoding moduleSparse 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 sourceDeepSeek-V4-Pro-0813 ↗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.

DeepSeek-V4-Pro-0813 42.7/100Rank 22/49

Survived 1 of 3
Foresight: complete playing field
Level 39 ptsTurn 20 / 20

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

LevelDeepSeek-V4-Pro-0813Step 5 Preview
Level 1Second Thoughts
14 ptsOut of lives
22 ptsSurvived
Level 2Pressure Pot
6 ptsOut of lives
Pending
Level 3Hide or Flight
9 ptsSurvived
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.

DeepSeek-V4-Pro-0813 71.7/100Rank 16/50

Cleared 5 of 5
Latch: complete playing field
Level 520.6 ptsTurn 27 / 40

Step 5 Preview 40.7/100Rank 28/50

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

Points per level

LevelDeepSeek-V4-Pro-0813Step 5 Preview
Level 1Hold the door
5.2 ptsCleared
5.2 ptsCleared
Level 2Crossfeed
9.8 ptsCleared
9 ptsCleared
Level 3Relay vault
18.8 ptsCleared
18.6 ptsCleared
Level 4Passing bay
17.2 ptsClearedFinished early · 1 beacon left
15.6 ptsClearedFinished early · 3 beacons left
Level 5Second thoughts
20.6 ptsCleared
15 ptsFinished early

Newton's Mirror

Same task instructions

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

The judgment

DeepSeek-V4-Pro-0813

The central sphere and gross paint state are coherent. Reduced circular side images and paired overhead quadrilaterals offer limited, spatially organized reflection cues, but the floor-sphere image, receding mirror-room structure, outer chains, nested imagery and curved emitter mappings are largely missing or invalid.

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

CategoryDeepSeek-V4-Pro-0813Step 5 Preview
State at 600 ms22.0/2520.0/25
Camera and placement15.0/1513.5/15
Reflections11.0/12020.0/120
Materials and scale8.0/107.0/10
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

DeepSeek-V4-Pro-0813

The whale’s warm eye, lavender assistant and tilted shell convey appealing character. However, the shrimp and reciprocal inspection are unresolved, the large bubble and upper tail are cropped, and the flipper’s grip is unclear. Broad competing color fields and conspicuous stepped edges weaken the flowing composition. Gradients suggest volume, but lighting and translucent materials remain uneven.

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

CategoryDeepSeek-V4-Pro-0813Step 5 Preview
Task & likeness3.0/106.0/10
Anatomy & construction5.0/104.5/10
Space & placement4.0/105.0/10
Light & reflections5.5/105.5/10
Purposeful detail5.5/105.5/10
Artistry & character2.5/105.0/10
Materials & effects5.0/106.0/10
Visual integrity3.0/105.0/10

Benchmark task

DeepSeek-V4-Pro-0813

A whale and a tiny shrimp examine one another through bubble lenses, with an octopus nearby. The scene tests interacting characters, differences in scale and transparent materials.

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

DeepSeek-V4-Pro-0813

Dark hair, a restrained smile and a hazy landscape make the reference recognizable. Severe construction problems dominate: oversized looping sleeves surround a background-colored opening, and the crossed hands lack coherent wrist and forearm connections. Rigid symmetry and elongated facial proportions weaken expression. Soft skin gradients and clean contours provide limited finish.

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

CategoryDeepSeek-V4-Pro-0813Step 5 Preview
Task & likeness2.0/104.0/10
Anatomy & construction0.5/104.0/10
Space & placement2.0/104.0/10
Light & reflections2.0/103.5/10
Purposeful detail2.0/104.0/10
Artistry & character3.0/104.0/10
Materials & effects1.5/103.0/10
Visual integrity1.0/103.5/10

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

DeepSeek-V4-Pro-0813

The glowing sunset, framing palms and layered shoreline create a coherent beach scene with animals, birds and playful accessories. The airborne blue animal has unclear anatomy, while flat foliage, weak ground contact and broad oval reflections limit depth and finish. The image is appealing but visibly simplified.

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

CategoryDeepSeek-V4-Pro-0813Step 5 Preview
Task & likeness6.5/104.0/10
Anatomy & construction4.0/103.0/10
Space & placement5.5/103.5/10
Light & reflections6.0/103.5/10
Purposeful detail3.5/103.0/10
Artistry & character4.0/103.0/10
Materials & effects6.0/103.5/10
Visual integrity5.0/104.0/10

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

DeepSeek-V4-Pro-0813

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

DeepSeek-V4-Pro-0813

The masked raccoon riding the steering dog is readily recognizable, with expressive faces and energetic diagonal movement. The oversized tail dominates the composition, limb connections are difficult to trace, and the jet ski has little structural depth. Blurred spray and a heavy oval shadow weaken water contact, while the glowing sun produces little corresponding illumination.

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

CategoryDeepSeek-V4-Pro-0813Step 5 Preview
Task & likeness7.0/104.0/10
Anatomy & construction4.5/104.5/10
Space & placement6.0/103.0/10
Light & reflections4.5/103.5/10
Purposeful detail6.0/104.0/10
Artistry & character4.5/103.5/10
Materials & effects5.0/103.0/10
Visual integrity4.0/103.0/10

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

DeepSeek-V4-Pro-0813

A clearly recognizable pelican cycles through a balanced, expressive illustration. The sweeping bill, shaded plumage and carefully drawn wheels give it character. Mechanical construction is weaker: the chain loop sits below the crank, and foot-to-pedal and wing-to-handlebar contact are ambiguous. Smooth gradients provide volume, but the scene remains visibly stylized.

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

CategoryDeepSeek-V4-Pro-0813Step 5 Preview
Task & likeness8.0/103.0/10
Anatomy & construction5.0/103.0/10
Space & placement6.5/103.0/10
Light & reflections7.0/103.0/10
Purposeful detail7.0/104.0/10
Artistry & character5.5/103.0/10
Materials & effects7.0/103.0/10
Visual integrity5.0/102.0/10

Benchmark task

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