GPT 6 Astra (Xhigh)
- Accuracy
- 85.4%
- Cost
- $11.54
- Time Spent
- 29m 52s
- Harness
- Codex · v0.160.1
- Tokens in / out
- 233.4K / 35.6K
- Date
3D view
Loading 3D model…
Drag the divider to reveal the build · drag the model to orbit · scroll to zoom
Model views
front
Loading front comparison images
back
Loading back comparison images
left
Loading left comparison images
right
Loading right comparison images
Details
- Parts placed
- 314
- Reference parts
- 314
- Parts matched
- 277
- Positions matched
- 265
- Orientations matched
- 247
- Surplus parts
- 0
- Match
- 0.838
- Bad parts
- 0
- Structural
- 1
- Car shape
- 1
Other builds of this benchmarkOther AI models, coding harnesses or reasoning effort levels on this same benchmark. Each row represents one setup and shows its build with the highest reference-accuracy score. Rows run from the highest score to the lowest. This build's setup is excluded.
| Harness · model | Accuracy | Cost | Time Spent |
|---|---|---|---|
| Codex · GPT 6 Astra (Ultra) | 84.9% | $22.19 | 1h 33m |
| Claude Code · Claude Sonnet 5.5 (Xhigh) | 70.2% | $135.43 | 3h 39m |
| Codex · GPT 6.1 Sol (Ultra) | 65.1% | $6.96 | 3h 47m |
| Codex · GPT 6.1 Sol (Xhigh) | 45.6% | $9.85 | 3h 59m |
| Claude Code · Claude Opus 5.5 (Xhigh) | 27.9% | $60.49 | 3h 10m |
| Claude Code · Claude Fable 5.1 (Max) | 19.0% | $89.40 | 2h 48m |
| Claude Code · Claude Haiku 5.5 (Xhigh) | 18.9% | $6.33 | 1h 53m |
| Claude Code · Claude Haiku 5.5 (Max) | 16.4% | $1.71 | 1h 9m |
| Grok Build · Grok 4.6 (Xhigh) | 12.4% | $10.77 | 58m 23s |
| Codex · GPT 6 Luna (Max) | 10.0% | $1.76 | 3h 49m |
