← mecheval / task / a3-cross-shaft-01

Two perpendicular cylinders fused into a cross A · A3 · a3-cross-shaft-01

rotation · boolean · steinmetz · advanced

Expected

Prompt

Make a single solid that is the union of two cylinders intersecting at right angles. Vertical cylinder: outside diameter 20mm, length 40mm, axis along Z, base on the XY plane (z = 0 to z = 40), centered on the Z axis. Horizontal cylinder: outside diameter 20mm, length 40mm, axis along the X axis, centered at (0, 0, 20) (so it passes through the middle of the vertical cylinder). The two cylinders are merged into one solid (their intersection region counts only once).

Checks

0
valid_solid
{
  "type": "valid_solid"
}
1
bbox
{
  "type": "bbox",
  "min": [
    -20,
    -10,
    0
  ],
  "max": [
    20,
    10,
    40
  ],
  "tolerance_mm": 0.2
}
2
mass_props
{
  "type": "mass_props",
  "volume_mm3": 19799.41,
  "tolerance_pct": 1.5
}
3
step_roundtrip
{
  "type": "step_roundtrip",
  "tolerance_pct": 1.5
}

Anti-cheese

{
  "max_solid_count": 1
}

Limits

{
  "max_tokens": 40000,
  "max_wallclock_sec": 240,
  "max_tool_calls": 40
}

Recent attempts

Runs (36)

modelrun statusscorefirst failtokenswall
claude-mcp-claude-opus-4-7 20260611T180747Z-f74d PASS 1.00 356.9k 116.5s
claude-mcp-claude-opus-4-7 20260611T180742Z-a345 fail 0.75 mass_props · volume off by 26.7% 585.2k 115.2s
claude-mcp-claude-opus-4-7 20260611T180643Z-71bd fail 0.50 bbox · Z off by +20.26mm 337.8k 59.3s
claude-mcp-claude-opus-4-7 20260611T180638Z-1980 PASS 1.00 632.5k 161.2s
claude-mcp-claude-opus-4-7 20260611T180618Z-d05c PASS 1.00 511.4k 89.4s
openai-direct-gpt-5-mini 20260429T001508Z-7980 fail 0.75 step_roundtrip · STEP drift on 1/1 solid 2.6k 39.8s
openai-direct-gpt-5-mini 20260429T001423Z-1f2b fail 0.75 step_roundtrip · STEP drift on 1/1 solid 2.4k 45.2s
openai-direct-gpt-5-mini 20260429T001336Z-3b34 fail 0.75 step_roundtrip · STEP drift on 1/1 solid 2.6k 46.9s
openai-direct-gpt-5-mini 20260429T001255Z-9522 fail 0.75 step_roundtrip · STEP drift on 1/1 solid 2.6k 41.2s
openai-direct-gpt-5-mini 20260429T001214Z-1863 fail 0.75 step_roundtrip · STEP drift on 1/1 solid 2.6k 40.6s
openai-direct-gpt-5 20260429T001154Z-944e fail 0.75 step_roundtrip · STEP drift on 1/1 solid 2.5k 22.7s
openai-direct-gpt-5 20260429T001130Z-432d fail 0.75 step_roundtrip · STEP drift on 1/1 solid 2.7k 23.8s
openai-direct-gpt-5 20260429T001050Z-cfb7 fail 0.75 step_roundtrip · STEP drift on 1/1 solid 2.4k 39.8s
openai-direct-gpt-5 20260429T001021Z-7fcd fail 0.75 step_roundtrip · STEP drift on 1/1 solid 3.0k 29.1s
openai-direct-gpt-5 20260429T000954Z-abff fail 0.75 step_roundtrip · STEP drift on 1/1 solid 2.5k 26.5s
openai-direct-gpt-5-mini 20260428T233319Z-c959 fail 0.75 step_roundtrip · STEP drift on 1/1 solid 2.3k 20.4s
openai-direct-gpt-4o-mini 20260428T232749Z-b807 fail 0.00 valid_solid · solid invalid 986 8.0s
openai-direct-gpt-4o-mini 20260428T232743Z-1d98 fail 0.00 valid_solid · solid invalid 919 5.8s
openai-direct-gpt-4o-mini 20260428T232738Z-32b0 fail 0.00 valid_solid · solid invalid 918 5.0s
claude-direct-claude-sonnet-4-6 20260428T232735Z-192c fail 0.75 step_roundtrip · STEP drift on 1/1 solid 1.2k 4.8s
openai-direct-gpt-4o-mini 20260428T232732Z-a87a fail 0.00 valid_solid · solid invalid 960 5.6s
claude-direct-claude-sonnet-4-6 20260428T232730Z-f8fb fail 0.75 step_roundtrip · STEP drift on 1/1 solid 1.2k 5.1s
openai-direct-gpt-4o-mini 20260428T232727Z-67ea fail 0.00 valid_solid · solid invalid 948 5.2s
claude-direct-claude-sonnet-4-6 20260428T232725Z-94fc fail 0.75 step_roundtrip · STEP drift on 1/1 solid 1.2k 5.3s
claude-direct-claude-sonnet-4-6 20260428T232720Z-9d98 fail 0.75 step_roundtrip · STEP drift on 1/1 solid 1.2k 4.7s
claude-direct-claude-sonnet-4-6 20260428T232715Z-bef3 fail 0.75 step_roundtrip · STEP drift on 1/1 solid 1.2k 4.4s
claude-direct-claude-haiku-4-5-20251001 20260428T232633Z-bdde fail 0.75 step_roundtrip · STEP drift on 1/1 solid 1.2k 2.1s
claude-direct-claude-haiku-4-5-20251001 20260428T232630Z-726a fail 0.00 valid_solid · solid invalid 1.3k 2.6s
claude-direct-claude-haiku-4-5-20251001 20260428T232628Z-7d7c fail 0.75 step_roundtrip · STEP drift on 1/1 solid 1.2k 2.6s
claude-direct-claude-haiku-4-5-20251001 20260428T232625Z-881a fail 0.00 valid_solid · solid invalid 1.2k 2.6s
claude-direct-claude-haiku-4-5-20251001 20260428T232619Z-68b0 fail 0.00 valid_solid · solid invalid 1.3k 5.6s
claude-direct-claude-opus-4-7 20260428T232020Z-9c1f fail 0.75 step_roundtrip · STEP drift on 1/1 solid 1.3k 3.8s
claude-direct-claude-opus-4-7 20260428T232016Z-cdfb fail 0.75 step_roundtrip · STEP drift on 1/1 solid 1.3k 4.1s
claude-direct-claude-opus-4-7 20260428T232011Z-6c4b fail 0.75 step_roundtrip · STEP drift on 1/1 solid 1.4k 4.1s
claude-direct-claude-opus-4-7 20260428T232006Z-dd44 fail 0.75 step_roundtrip · STEP drift on 1/1 solid 1.4k 4.7s
claude-direct-claude-opus-4-7 20260428T232002Z-8cb9 fail 0.75 step_roundtrip · STEP drift on 1/1 solid 1.4k 4.5s

generated 2026-06-17T03:16:07.223Z · static site, regenerate with npm run build -w @mecheval/leaderboard