BrickAGI benchmark

models /

claude-opus-4-8-coordinator-blind scaffold-assembly-v1

Provider: anthropic · Run at 2026-06-08 · Submission b1a9c0de

Build Points 382 / 2,631 available
CaSS 57.4%
CBS 30.3%
Coverage 51%
Cost $0
Tokens (in/out) 0 / 0
Diagnostic row, not ranked. legacy task-set provenance

Tier breakdown

The 3-dimension diagnostic matrix for this submission. A tier shows no CBS until the structural prover returns a definitive verdict there.

Tier BOM pass Scope pass Buildability CaSS CBS Coverage n
trivial 100.0% 100.0% 75.0% 100.0% 75.0% 100% 4
easy 100.0% 100.0% 66.7% 100.0% 50.0% 75% 4
medium 44.4% 44.4% 75.0% 44.4% 33.3% 44% 9
hard 57.1% 57.1% 50.0% 57.1% 28.6% 57% 7
stretch 100.0% 100.0% 25.0% 100.0% 25.0% 100% 4
expert 0.0% 0.0% — 0.0% 0.0% 0% 4
master 0.0% 0.0% — 0.0% 0.0% 0% 5

Per-task results

All 37 tasks attempted by this baseline. Click a task to see the same row from every model's perspective.

# Slug Tier CaSS BOM Scope Build Ref sim Iter Tokens Duration Status 3D
0001 flat-plate-5x7 trivial 1 1 1 1 1.000 1 0 0ms pass 3D
0002 flat-plate-5x5 trivial 1 1 1 1 1.000 1 0 0ms pass 3D
0003 staggered-column-2x2 trivial 1 1 1 1 1.000 1 0 0ms pass 3D
0004 hollow-frame-8x8 trivial 1 1 1 0 1.000 1 0 0ms pass 3D
0005 small-table easy 1 1 1 1 0.000 1 0 0ms pass 3D
0006 step-pyramid easy 1 1 1 1 0.400 1 0 0ms pass 3D
0007 simple-gate easy 1 1 1 0 1.000 1 0 0ms pass 3D
0008 picture-frame easy 1 1 1 incon. 0.412 1 0 0ms pass 3D
0009 minifig-chair medium 1 1 1 1 1.000 1 0 0ms pass 3D
0010 bookshelf-3-shelf medium 1 1 1 0 0.118 1 0 0ms pass 3D
0011 simple-bridge medium 1 1 1 1 1.000 1 0 0ms pass 3D
0012 square-lantern medium 1 1 1 1 1.000 1 0 0ms pass 3D
0013 farm-tractor hard 1 1 1 1 0.400 1 0 0ms pass 3D
0014 rolling-cart hard 1 1 1 0 0.379 1 0 0ms pass 3D
0015 drawbridge hard 1 1 1 1 1.000 1 0 0ms pass 3D
0016 slatted-bench hard 1 1 1 0 0.540 1 0 0ms pass 3D
0017 small-house stretch 1 1 1 0 0.236 1 0 0ms pass 3D
0018 small-lighthouse stretch 1 1 1 1 0.126 1 0 0ms pass 3D
0019 garden-shed stretch 1 1 1 0 0.025 1 0 0ms pass 3D
0020 fountain-with-pool stretch 1 1 1 0 0.173 1 0 0ms pass 3D
0021 grand-bridge expert 0 0 0 incon. 0.000 1 0 0ms DNC —
0022 manor-house expert 0 0 0 incon. 0.000 1 0 0ms DNC —
0023 clock-tower expert 0 0 0 incon. 0.000 1 0 0ms DNC —
0024 grand-fountain expert 0 0 0 incon. 0.000 1 0 0ms DNC —
0025 mosaic-smiley master 0 0 0 incon. 0.000 1 0 0ms DNC —
0026 voxel-tree master 0 0 0 incon. 0.000 1 0 0ms DNC —
0027 grand-mosaic-sunset master 0 0 0 incon. 0.000 1 0 0ms DNC —
0028 mega-voxel-rocket master 0 0 0 incon. 0.000 1 0 0ms DNC —
0029 colossal-mosaic-grid master 0 0 0 incon. 0.000 1 0 0ms DNC —
0030 bonded-plate-7x7 medium 0 0 0 incon. 0.000 1 0 0ms DNC —
0031 bonded-platform-9x5 medium 0 0 0 incon. 0.000 1 0 0ms DNC —
0032 frame-6x6 medium 0 0 0 incon. 0.000 1 0 0ms DNC —
0033 grand-table-6x6 medium 0 0 0 incon. 0.000 1 0 0ms DNC —
0034 great-pyramid-8x8 medium 0 0 0 incon. 0.000 1 0 0ms DNC —
0035 wide-gate hard 0 0 0 incon. 0.000 1 0 0ms DNC —
0036 throne hard 0 0 0 incon. 0.000 1 0 0ms DNC —
0037 tall-column-2x2 hard 0 0 0 incon. 0.000 1 0 0ms DNC —

Submitted BOMs

The final BOM the model emitted on each task it attempted.

#0001 flat-plate-5x7 trivial CaSS 1

5x7 flat red plate (bonded two-layer)
part_num color_id qty
3032 4 2
3666 4 2
3710 4 2
3024 4 2
total pieces 8

#0002 flat-plate-5x5 trivial CaSS 1

5x5 flat red plate (bonded two-layer)
part_num color_id qty
78329 4 10
total pieces 10

#0003 staggered-column-2x2 trivial CaSS 1

Staggered 2x2 column, 13 bricks tall (white)
part_num color_id qty
3004 15 26
total pieces 26

#0004 hollow-frame-8x8 trivial CaSS 1

Hollow square frame 8x8 (6x6 hole, corner-bonded)
part_num color_id qty
3460 15 2
3666 15 2
3023 15 4
total pieces 8

#0005 small-table easy CaSS 1

Small LEGO table (4x4 top on four 3-brick legs)
part_num color_id qty
3031 19 1
3005 70 12
total pieces 13

#0006 step-pyramid easy CaSS 1

Tan step pyramid (6x6 / 4x4 / 2x2, three stacked plate layers)
part_num color_id qty
3958 19 1
3031 19 1
3022 19 1
total pieces 3

#0007 simple-gate easy CaSS 1

Simple LEGO gate: two 5-brick pillars 4 studs apart bridged by a 1x6 lintel
part_num color_id qty
3005 15 10
3666 15 1
total pieces 11

#0008 picture-frame easy CaSS 1

Standing picture frame (brown frame, white 4x6 picture, base foot)
part_num color_id qty
3035 70 1
3032 15 1
3010 70 6
3009 70 2
3622 70 2
total pieces 12

#0009 minifig-chair medium CaSS 1

Minifig chair (2x2 red seat, four 2-brick legs, brick-built backrest)
part_num color_id qty
3022 4 1
3005 4 8
3004 4 3
total pieces 12

#0010 bookshelf-3-shelf medium CaSS 1

Brown 3-shelf bookshelf (2 side walls, closed back)
part_num color_id qty
3010 70 18
3035 70 3
3666 70 6
total pieces 27

#0011 simple-bridge medium CaSS 1

Simple stone-grey footbridge: two 2x2 pillars, bonded 4x12 deck, railings
part_num color_id qty
3003 71 8
3035 71 1
3020 71 2
3032 71 2
3005 71 12
total pieces 25

#0012 square-lantern medium CaSS 1

Square Lantern (4x4, dark bluish gray)
part_num color_id qty
3031 72 2
3010 72 8
3005 72 4
total pieces 14

#0013 farm-tractor hard CaSS 1

Green farm tractor (chassis, engine, open cab, 4 wheels on holders)
part_num color_id qty
3035 2 1
30157 2 2
56145 0 2
92402 0 2
4624 0 2
3641 0 2
3008 2 2
3003 2 2
3022 2 2
3020 2 2
4079 2 1
3062b 2 1
total pieces 21

#0014 rolling-cart hard CaSS 1

Rolling cart with four wheels and front hitch
part_num color_id qty
3035 70 1
30157 70 2
4624 71 4
3641 0 4
3008 70 4
3004 70 4
4623 70 1
total pieces 20

#0015 drawbridge hard CaSS 1

Castle drawbridge: two 5-brick stone pillars 4 studs apart, brown bridge plate on a floor-mounted hinge
part_num color_id qty
3004 71 10
3795 71 1
3937 71 1
3830 70 1
3710 70 1
3020 70 1
total pieces 15

#0016 slatted-bench hard CaSS 1

Slatted park bench (12x4 seat, perpendicular bonding, slatted backrest)
part_num color_id qty
3010 70 9
3710 70 14
60479 70 10
60479 308 2
30414 70 2
3004 70 6
3062b 308 4
3024 70 12
total pieces 59

#0017 small-house stretch CaSS 1

Small LEGO House (white walls, brown door, red sloped roof)
part_num color_id qty
92438 71 1
3008 15 22
3009 15 1
3622 15 2
3004 15 12
3005 15 14
60596 70 1
60616 70 1
60593 15 1
3023 15 72
3034 4 4
3020 4 2
3037 4 8
3039 4 4
3795 4 4
3460 4 2
total pieces 151

#0018 small-lighthouse stretch CaSS 1

Small Lighthouse
part_num color_id qty
3021 71 48
3031 71 1
3009 71 20
3010 71 20
3666 71 2
3710 71 2
3941 4 6
3941 15 6
3031 4 1
3062b 15 4
3024 15 8
3005 47 4
3022 47 1
6233 4 1
total pieces 124

#0019 garden-shed stretch CaSS 1

Garden Shed (dark tan walls, dark grey shed roof, brown door)
part_num color_id qty
41539 72 1
3008 69 9
3009 69 8
3622 69 2
3004 69 6
3005 69 6
60596 70 1
60616 70 1
60593 69 1
92438 72 1
3037 72 4
6065 28 1
3024 72 104
total pieces 145

#0020 fountain-with-pool stretch CaSS 1

Garden fountain with 10x10 pool, perimeter wall, central round pedestal and basin
part_num color_id qty
3030 71 2
3033 71 2
3035 71 2
3832 71 1
3710 71 12
3010 71 8
3004 71 2
3023 71 2
3024 71 36
3024 43 39
3022 43 8
3460 43 2
3666 43 2
3941 71 4
4032b 71 1
3062b 71 1
3062b 43 6
total pieces 130

#0021 grand-bridge DNC

no solution produced

#0022 manor-house DNC

no solution produced

#0023 clock-tower DNC

no solution produced

#0024 grand-fountain DNC

no solution produced

#0025 mosaic-smiley DNC

no solution produced

#0026 voxel-tree DNC

no solution produced

#0027 grand-mosaic-sunset DNC

no solution produced

#0028 mega-voxel-rocket DNC

no solution produced

#0029 colossal-mosaic-grid DNC

no solution produced

#0030 bonded-plate-7x7 DNC

no solution produced

#0031 bonded-platform-9x5 DNC

no solution produced

#0032 frame-6x6 DNC

no solution produced

#0033 grand-table-6x6 DNC

no solution produced

#0034 great-pyramid-8x8 DNC

no solution produced

#0035 wide-gate DNC

no solution produced

#0036 throne DNC

no solution produced

#0037 tall-column-2x2 DNC

no solution produced

Run metadata

Submission IDb1a9c0de-1f43-4ddd-8aaa-coordblind14
Protocolscaffold-assembly-v1
Protocol hashsha256:394f2d859fafd42d0ac144c5f251e68ea2e7f4c06da48bf733e1c2505055ee9c
Run at2026-06-08T00:00:00Z
Total duration0ms
Adapter healthn/a