STEALTHMARK · HEAD TO HEAD

All comparisons ↗

DeepSeek-V4.1-Flash-1017 vs GPT-5.6 Luna

DeepSeek-V4.1-Flash-1017 leads GPT-5.6 Luna by 21.4 points on StealthMark, 59.2 to 37.8. It leads on all 9 benchmarks. DeepSeek-V4.1-Flash-1017 scores higher on Newton's Mirror, with its largest category advantage in reflections.

See the difference ↘
All results ↗
All results ↗

Overall StealthMark

59.2/100Overall rank 9/53
37.8/100Overall rank 37/53

General Intelligence

56.8/100
30.4/100

Visual Intelligence

63.5/100
50.2/100

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

  • GPT-5.6 Luna

Open or local

Open modelMIT

BENCHMARK BY BENCHMARK

Where each model leads

  • DeepSeek-V4.1-Flash-1017
  • GPT-5.6 Luna
  • Other models
  • Median

DATASHEET

Specs and pricing

DeepSeek-V4.1-Flash-1017 scores higher on StealthMark and runs 1.7× faster in our speed test. Both take about 1M tokens of context.

  • DeepSeek-V4.1-Flash-1017
  • GPT-5.6 Luna
  • Other models
  • Median
PricePer 1M output tokens, input below
DeepSeek-V4.1-Flash-1017: $1.20Tied$0.30 inputGPT-5.6 Luna: $1.20Tied$0.20 input
Context windowTokens per request, max output below
DeepSeek-V4.1-Flash-1017: 1MTied384K max outputGPT-5.6 Luna: 1.05MTied128K max output
Output speedTokens per second, our measurement
DeepSeek-V4.1-Flash-1017: 3101.7× fastertokens/sGPT-5.6 Luna: 181tokens/s
SizeTotal and active parameters
DeepSeek-V4.1-Flash-1017: 763B16B activeEstimatedGPT-5.6 Luna: Not disclosed
Developer
DeepSeek-AIOpenAI
Released
17 Oct 20269 Jul 2026
Origin
ChinaUnited States
Input → output
Text, image → textBoth
Architecture
Sparse mixture-of-experts with a causal encoder-decoder design and native image understanding.Not published

Full datasheets, every price and sourceDeepSeek-V4.1-Flash-1017 ↗GPT-5.6 Luna ↗

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-V4.1-Flash-1017 52.9/100Rank 15/47

Survived 1 of 3
Foresight: complete playing field
Level 310 ptsTurn 15 / 20

GPT-5.6 Luna 25.1/100Rank 35/47

Survived 0 of 3
Foresight: complete playing field
Level 38 ptsTurn 15 / 20

Points per level

LevelDeepSeek-V4.1-Flash-1017GPT-5.6 Luna
Level 1Second Thoughts
26 ptsSurvived
4 ptsOut of lives
Level 2Pressure Pot
7 ptsOut of lives
6 ptsOut of lives
Level 3Hide or Flight
10 ptsOut of lives
8 ptsOut of lives

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-V4.1-Flash-1017 73.9/100Rank 13/48

Cleared 5 of 5
Latch: complete playing field
Level 520 ptsTurn 30 / 40

GPT-5.6 Luna 52.9/100Rank 21/48

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

Points per level

LevelDeepSeek-V4.1-Flash-1017GPT-5.6 Luna
Level 1Hold the door
5.2 ptsCleared
5.2 ptsCleared
Level 2Crossfeed
9.8 ptsCleared
8.6 ptsClearedFinished early · 1 beacon left
Level 3Relay vault
19.2 ptsCleared
18 ptsCleared
Level 4Passing bay
16.4 ptsClearedFinished early · 3 beacons left
17 ptsClearedFinished early · 3 beacons left
Level 5Second thoughts
20 ptsCleared
20.6 ptsCleared

Newton's Mirror

Same task instructions

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

The judgment

DeepSeek-V4.1-Flash-1017

Principal sphere and airborne-black placements are supported, and diminishing bilateral chains provide limited optical structure. Missing nested imagery, incomplete outer chains, distorted first images and sparse emitter recurrence remain the dominant weaknesses. Oversized paint bodies and unresolved cyan contact impose separate limitations.

GPT-5.6 Luna

The room enclosure has coherent symmetric perspective, with the overhead panel, sphere and black aligned centrally. The sphere and airborne black occupy substantially lower positions than the reference framing, inconsistent with the prescribed camera aim. The diagonal crimson streak does not establish the required surface transport. Faint wall sphere images provide limited mirrored correspondence, but the shallow floor image, missing receding chains, absent nested sphere imagery and sparse emitter mapping leave most reflection structure unsupported.

Score breakdown

CategoryDeepSeek-V4.1-Flash-1017GPT-5.6 Luna
State at 600 ms23.0/2515.0/25
Camera and placement13.5/154.0/15
Reflections34.0/1203.5/120
Materials and scale8.5/106.5/10
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-V4.1-Flash-1017

An expressive, intricately drawn storybook scene with a complete flowing silhouette and differentiated glossy materials. The supported main lens and continuous shrimp body make the central interaction readable, but reciprocal eye lines, facial magnification and the smaller lens grip remain ambiguous. Bright granular fringes and crowded accessory lines limit the finish.

GPT-5.6 Luna

Six pale, outward-facing heads and six rounded lobes make the knot recognizable. The central hexagonal opening remains clear. However, broad gray crossing panels flatten the weave, several inner joins resemble capped strips, and sparse oversized scale marks leave the bodies looking more like decorated ribbons than living snakes.

Score breakdown

CategoryDeepSeek-V4.1-Flash-1017GPT-5.6 Luna
Task & likeness7.5/107.5/10
Anatomy & construction7.5/106.0/10
Space & placement8.0/106.5/10
Light & reflections7.0/105.0/10
Purposeful detail7.5/105.5/10
Artistry & character8.0/106.0/10
Materials & effects7.5/105.0/10
Visual integrity6.5/106.0/10

Benchmark task

DeepSeek-V4.1-Flash-1017

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.

GPT-5.6 Luna

A pale serpent woven into an interlocking knot tests connected anatomy, repeated structure and readable overlaps. Organic surface detail must work with the overall geometry.

Mona Lisa

Different task instructions

Precise visual recall and portrait reconstruction.

The judgment

DeepSeek-V4.1-Flash-1017

A coherent graphic interpretation with controlled contours, modeled facial volume and purposeful portrait details. Replication remains limited by the frontal pose, enlarged face, altered expression and ambiguous crossed hands. Landscape layering provides depth, while schematic fabric, smooth surfaces and tonal bands constrain the finish.

GPT-5.6 Luna

The central portrait, dark clothing, crossed hands and muted landscape clearly evoke Mona Lisa. Replication is limited by the elongated face, conspicuous eyes and simplified smile. Overlapping fingers have uncertain anatomy, while broad gradients and heavy outlines flatten the portrait. The composition and restrained palette remain coherent.

Score breakdown

CategoryDeepSeek-V4.1-Flash-1017GPT-5.6 Luna
Task & likeness5.0/103.5/10
Anatomy & construction5.0/103.5/10
Space & placement6.0/104.0/10
Light & reflections6.0/104.5/10
Purposeful detail5.5/104.0/10
Artistry & character5.5/104.0/10
Materials & effects5.5/103.5/10
Visual integrity6.0/104.5/10

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-V4.1-Flash-1017

A coherent, colorful beach illustration with controlled curves, clean layering and varied readable accessories. Castle details and small animal features strengthen its finish. Schematic airborne animals, uncertain board support, shallow architectural depth and decorative reflections limit physical coherence and drawing sophistication.

GPT-5.6 Luna

The warm sunset, palm silhouette and varied beach activities create an appealing, readable composition. Animals, birds and a broad sun reflection satisfy the brief. Drawing remains elementary: angular airborne sea animals lack splash or convincing motion, the crab's claws are awkwardly constructed, and repeated reflection bands suggest a stepped platform. Clean outlines and pleasant colors exceed the anatomical and spatial sophistication.

Score breakdown

CategoryDeepSeek-V4.1-Flash-1017GPT-5.6 Luna
Task & likeness7.0/107.5/10
Anatomy & construction5.5/104.0/10
Space & placement5.5/105.5/10
Light & reflections6.0/104.0/10
Purposeful detail6.5/106.5/10
Artistry & character5.0/103.5/10
Materials & effects5.5/104.0/10
Visual integrity4.5/103.5/10

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-V4.1-Flash-1017

This benchmark has not been assessed yet.

GPT-5.6 Luna

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

Different task instructions

Scene composition and interacting subjects.

The judgment

DeepSeek-V4.1-Flash-1017

An appealing cartoon with recognizable riders, expressive faces and a cohesive twilight palette. Clean shapes support the composition, but ambiguous limb overlaps, weak seat contact and loosely resolved steering limit physical coherence. Lighting and water effects are decorative, with modest material differentiation and drawing complexity.

GPT-5.6 Luna

Expressive animals, a striped tail and clean outlines make the scene appealing and readable. The raccoon sits convincingly on the dog's back, but the overlapping steering paws are ambiguous and the control stem stops above the hull. Broad wake shapes suggest motion; the glowing sun lacks a corresponding water reflection.

Score breakdown

CategoryDeepSeek-V4.1-Flash-1017GPT-5.6 Luna
Task & likeness8.0/107.0/10
Anatomy & construction7.0/105.0/10
Space & placement7.0/106.5/10
Light & reflections6.5/106.0/10
Purposeful detail7.0/106.5/10
Artistry & character7.0/104.5/10
Materials & effects6.5/105.5/10
Visual integrity8.0/105.0/10

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-V4.1-Flash-1017

An expressive, controlled illustration with continuous body construction, substantial bicycle detail and effective landscape depth. Unclear foot-to-pedal and wing-to-handlebar contacts limit physical conviction and local readability. Purposeful detail and composition are stronger than directional lighting and material treatment.

GPT-5.6 Luna

A recognizable pelican cyclist with an expressive profile, balanced composition and controlled outlines. Head, neck and torso connect coherently. The bicycle wheels and frame are convincing, but overlapping feet obscure pedal contact, the drivetrain lacks a clear chain, and the wing's handlebar grip is schematic. Atmospheric color succeeds more strongly than directional lighting or material depth.

Score breakdown

CategoryDeepSeek-V4.1-Flash-1017GPT-5.6 Luna
Task & likeness8.5/108.5/10
Anatomy & construction7.0/105.5/10
Space & placement7.5/107.5/10
Light & reflections6.5/106.5/10
Purposeful detail7.5/106.5/10
Artistry & character7.0/105.5/10
Materials & effects7.0/107.0/10
Visual integrity7.5/106.5/10

Benchmark task

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