partforge

Photo knob

hard tier · create · with a reference image

The request

The attached photo shows a knurled knob from the side, standing on its base. It is 38.13 mm across at its widest. In the top face, which the photo cannot show, there is a blind bore: 9.5 mm in diameter, 10.64 mm deep, with a flat on one side of the bore 8.58 mm from the opposite wall, the flat facing +Y. Make one that matches the picture: the same proportions, the chamfers at both rims, the knurl, and that bore. Z up, the knob standing on Z = 0 with its axis on the origin. Name the sub-part `knob`.
The photograph the model was given
Photo · made with iPhone camera · CC0-1.0

What the AI judge checks

  • The overall form matches the reference: the same kind of object, the same silhouette from the matching angle
  • Proportions match: the relative sizes of the major dimensions agree with the reference within what the eye can judge
  • Every major feature visible in the reference (holes, slots, bosses, tabs, lips) is present in the result
  • Those features sit where the reference has them
  • Nothing is present that the reference does not show — no decorative or unexplained geometry
  • A diamond knurl around the side — crossing grooves, not smooth and not straight ridges
  • A flat top face with a single hole in its centre and nothing else cut into it

Measured checks

  • watertight (gate)
  • one sub-part named knob (gate)
  • diameter is 38.13 mm
  • height is 12 mm ± 10 %
  • the bore clears 9.5 mm to 10.64 mm deep
  • the bore is not oversize
  • the D flat is where the prompt put it
  • the flat is not cut too deep
  • the bore is blind
  • chamfered at the top rim
  • chamfered at the bottom rim
  • the side is full at mid-height
  • the top face is flat out to 34.7 mm

Ranked by hand

One reviewer put one attempt from each model in order, best first, ties left tied.

  1. #1 tie
    Gemini 3.8 Flash
    Gemini 3.8 Flash's ranked attempt
  2. #1 tie
    GPT-5.6 Sol
    GPT-5.6 Sol's ranked attempt
  3. #3
    GPT-6 Astra
    GPT-6 Astra's ranked attempt
  4. #4
    Opus 5
    Opus 5's ranked attempt
  5. #5
    Fable 5.1
    Fable 5.1's ranked attempt

Ordering by hand decides Human rank and Human rate on the leaderboard; it never affects Overall or Strict.

Every model's attempts

Run baseline-2026-092026-09-10commit 41aec7daspend $339.70 + $48.02 AI judging, 7 unpricedOther runs: 2026-09-04T20-02-10-stage2-hard-cases, 2026-09-04T19-48-31-screen-14-fixed