STEALTHMARK · HEAD TO HEAD

All comparisons ↗

GPT-6 Luna vs LongCat-2.0

GPT-6 Luna leads LongCat-2.0 by 13.6 points on StealthMark, 39.1 to 25.5. It leads on 5 of the 7 benchmarks with results for both and LongCat-2.0 on 2. GPT-6 Luna scores higher on Ground Truth.

See the difference ↘
All results ↗
Model guide ↗

Overall StealthMark

39.1/100Overall rank 40/58
25.5/100Overall rank 50/58

General Intelligence

31.3/100
17.2/100

Visual Intelligence

52.0/100
34.4/100

Reasoning efficiency

How many reasoning tokens a model needs for the quality it delivers.

Single-step reasoning How much it reasons to understand a task

Multi-step reasoning How much it keeps reasoning as a task goes on

  • LongCat-2.0Multi-step reasoning not measured

Cost efficiency

How much Performance you get for your money.

Open or local

Open modelMIT

BENCHMARK BY BENCHMARK

Where each model leads

  • GPT-6 Luna
  • LongCat-2.0
  • Other models
  • Median

DATASHEET

Specs and pricing

GPT-6 Luna scores higher on StealthMark, is 2.4× cheaper per output token and runs 1.5× faster in our speed test. Both take about 1.05M tokens of context.

  • GPT-6 Luna
  • LongCat-2.0
  • Other models
  • Median
PricePer 1M output tokens, input below
GPT-6 Luna: $0.502.4× cheaper$0.10 inputLongCat-2.0: $1.20$0.30 input
Context windowTokens per request, max output below
GPT-6 Luna: 1.05MTied128K max outputLongCat-2.0: 1MTied
Output speedTokens per second, our measurement
GPT-6 Luna: 1311.5× fastertokens/sLongCat-2.0: 86tokens/s
Developer
OpenAIMeituan
Released
22 Sep 2026Not published
Origin
United StatesChina
Input → output
Text, image → textText → text
Architecture
Not publishedMixture-of-Experts with LongCat Sparse Attention

Full datasheets, every price and sourceGPT-6 Luna ↗LongCat-2.0 ↗

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.

GPT-6 Luna 29.8/100Rank 37/54

Survived 0 of 3
Foresight: complete playing field
Level 36 ptsTurn 7 / 20

Points per level

LevelGPT-6 LunaLongCat-2.0
Level 1Second Thoughts
15 ptsOut of lives
No run
Level 2Pressure Pot
5 ptsOut of lives
No run
Level 3Hide or Flight
6 ptsOut of lives
No run

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.

GPT-6 Luna 44.0/100Rank 31/55

Cleared 4 of 5
Latch: complete playing field
Level 513 ptsTurn 15 / 40

Points per level

LevelGPT-6 LunaLongCat-2.0
Level 1Hold the door
5.2 ptsCleared
No run
Level 2Crossfeed
8.8 ptsClearedFinished early · 1 beacon left
No run
Level 3Relay vault
17 ptsCleared1 beacon left
No run
Level 4Passing bay
15.8 ptsClearedFinished early · 3 beacons left
No run
Level 5Second thoughts
13 ptsFinished early
No run

Newton's Mirror

Same task instructions

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

The judgment

GPT-6 Luna

The schematic scene supports the paint contact order and compact paint scale. Its low, undersized main sphere and incorrectly placed wall images depart from the reference geometry. A few shrinking wall images provide limited optical organization, while absent lateral recursion and outer chains, broad sphere ribbons and isolated panel highlights leave most nested room and emitter mapping undeveloped.

LongCat-2.0

Cyan lies on the sphere's facing middle rather than showing its recent interaction near the apex. Separate airborne black and elongated crimson support part of the timing sequence. Small side images and an overlapping lower image provide limited reflection ordering, but recursive room structure, nested imagery and emitter mappings are largely absent. The main sphere is smaller and higher in the frame than prescribed, and the surfaces have a soft, matte appearance.

Score breakdown

CategoryGPT-6 LunaLongCat-2.0
State at 600 ms22.0/259.5/25
Camera and placement3.5/156.5/15
Reflections5.0/12012.5/120
Materials and scale7.0/105.5/10
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

GPT-6 Luna

Sixfold symmetry, pale coloring and six outward-facing heads establish the motif. However, every neck floats clear of the knot. Squared outer sections replace the required semicircular lobes, and the large stepped opening lacks a regular hexagonal contour. Repeated scales and soft shading suggest skin, but translucent crossings and doubled edges weaken the living, interwoven construction.

LongCat-2.0

Amber slit eyes, a pointed beak, cream feathers and gold-trimmed black clothing establish the secretive owl rogue. The head appears frontal rather than lowered, with much of the face exposed. The centered silhouette is effective, but stiff cloak geometry, shallow facial modeling and extremely dark clothing limit depth and finish. Sparse gold markings suggest the requested circuitry only loosely.

Score breakdown

CategoryGPT-6 LunaLongCat-2.0
Task & likeness4.5/107.0/10
Anatomy & construction3.5/106.0/10
Space & placement5.5/106.0/10
Light & reflections5.0/105.5/10
Purposeful detail5.0/106.0/10
Artistry & character5.0/106.5/10
Materials & effects5.0/105.5/10
Visual integrity4.5/106.5/10

Benchmark task

GPT-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.

LongCat-2.0

A cloaked owl with luminous eyes tests expression, fabric folds and the separation of dark forms.

Mona Lisa

Different task instructions

Precise visual recall and portrait reconstruction.

The judgment

GPT-6 Luna

The dark veil, crossed arms, restrained smile and layered landscape clearly evoke Mona Lisa. Clean contours and muted colors create a coherent illustration. Replication is limited by enlarged, outlined eyes, simplified facial volume, broad shoulders and indistinct overlapping hands. Gradients suggest depth, but lack the original's subtle painterly transitions.

LongCat-2.0

The centered portrait, slight smile, overlapping hands and distant landscape suggest Mona Lisa, but replication is weak. The face has little volume, the shoulders and torso form a narrow wedge, and the hands are fingerless ovals with unresolved wrists. Broad gray background bands provide limited depth. A bright head rim and pale hands contrast with an almost featureless dark figure.

Score breakdown

CategoryGPT-6 LunaLongCat-2.0
Task & likeness4.5/100.5/10
Anatomy & construction4.0/100.5/10
Space & placement5.5/100.5/10
Light & reflections5.0/100.5/10
Purposeful detail5.0/100.0/10
Artistry & character4.5/100.5/10
Materials & effects5.0/100.0/10
Visual integrity6.0/100.5/10

Benchmark task

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

Beach

Different task instructions

Creativity, perspective and visual beauty.

The judgment

GPT-6 Luna

The warm sky, layered sea and framing palms create an appealing beach composition. Animals, birds, toys and reflected sunlight fulfill the brief. Human anatomy remains schematic, and simplified figures offer limited expression. Missing shadows and the dolphin's splashless jump weaken physical grounding, while clean silhouettes and readable small details support the illustration.

LongCat-2.0

The sunset, palms, birds and beach animals create a cheerful, readable scene. Clean silhouettes and a vivid sky establish the composition, but construction remains schematic. The right palm appears rooted in open water, shadows have limited connection to the sunset, and the sun’s reflection consists of rigid bands. Small creatures and beach equipment are recognizable but minimally developed.

Score breakdown

CategoryGPT-6 LunaLongCat-2.0
Task & likeness7.5/107.5/10
Anatomy & construction5.0/105.5/10
Space & placement7.5/105.5/10
Light & reflections5.0/106.0/10
Purposeful detail7.0/105.0/10
Artistry & character5.0/105.0/10
Materials & effects6.5/103.0/10
Visual integrity6.0/104.0/10

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

GPT-6 Luna

This benchmark has not been assessed yet.

LongCat-2.0

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

GPT-6 Luna

Expressive faces, warm sunset colors and clean silhouettes create an appealing illustration. The raccoon sits on the dog's back, but the dog's grip and steering linkage remain unclear. The broad hull resembles a speedboat, and flat shading, absent sun reflections and modest water displacement limit depth and motion.

LongCat-2.0

The raccoon is recognizable and expressive, with a clear mask and striped tail. The dog's anatomy is poorly resolved, and the flattened craft lacks convincing steering and seating. Feet extend below the hull, weakening the sense of support. An appealing palette, clean shapes and decorative water effects give the illustration some polish.

Score breakdown

CategoryGPT-6 LunaLongCat-2.0
Task & likeness8.0/104.0/10
Anatomy & construction5.5/103.0/10
Space & placement7.5/104.5/10
Light & reflections6.0/105.5/10
Purposeful detail7.0/106.0/10
Artistry & character5.5/103.0/10
Materials & effects6.5/105.5/10
Visual integrity5.0/105.0/10

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

Different task instructions

Subject construction and spatial relationships.

The judgment

GPT-6 Luna

The pelican’s distinctive bill and the coordinated vintage poster palette make an appealing illustration. Circular wheels, orderly spokes and clean lettering strengthen the design. Riding mechanics are less convincing: neither wing reaches the handlebars, pedal contact is ambiguous, and the drivetrain is incomplete. The wheels sit against water without a clear supporting surface. Flat lighting limits volume.

LongCat-2.0

A clearly recognizable pelican cycles through a well-composed coastal sunset. Warm colors, flowing scarf, feather contours and bicycle details give the illustration character. The legs connect to the body, but the farther foot's pedal contact and the wing's handlebar grip remain ambiguous. Clean vector shapes provide polish, with limited material texture and directional shading.

Score breakdown

CategoryGPT-6 LunaLongCat-2.0
Task & likeness8.0/108.5/10
Anatomy & construction5.0/107.0/10
Space & placement6.5/108.0/10
Light & reflections6.5/107.5/10
Purposeful detail7.5/108.0/10
Artistry & character6.5/108.5/10
Materials & effects6.5/107.0/10
Visual integrity7.0/108.5/10

Benchmark task

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