Model comparison
pick_cube_lora_1.8w vs o3
Capability radar, value bars, and side-by-side posture.
pick_cube_lora_1.8w
Not scored
value score
o3
93
value score
pick_cube_lora_1.8w ctx
—
o3 ctx
200K
Capability radar
Value · context · multimodal · openness · speed posture
Value head-to-head
- o393
Attribute tape
| Field | pick_cube_lora_1.8w | o3 |
|---|---|---|
| Developer | KaiyueChen | OpenAI |
| Context | — | 200K |
| Modalities | Text, Code | |
| Openness | — | Proprietary Model |
| Speed | — | — |
| Price | — | $10/1M in · $40/1M out |
| Value score | — | 93 |
| Summary | A LoRA (Low-Rank Adaptation) fine-tuning adapter checkpoint published by KaiyueChen on Hugging Face, targeting a 'pick cube' robotic manipulation task. The adapter adds task-specific weights on top of a frozen base model using parameter-efficient fine-tuning, with approximately 1.8 thousand training steps indicated by the suffix. | OpenAI reasoning model focused on hard problems. |