LetTokenBurnDVerifierBench Grounding Generative Verification in 3D ContentShuanglin Li, Jiaying Lin, Shuquan Ye.
$ npm i -g lettokenburn-cli$ ltb bench pull 3dverifierbench --n 50$ ltb bench run ./3dverifierbench -a claude-code -m anthropic/claude-opus-4-8$ ltb bench push
License is link-only: ltb bench pull fetches rows from the original host for local evaluation only and does not write a solution/ directory. Do not redistribute the generated tasks.
Runs happen on your machine with your own subscription; only the score summary is pushed. CLI docs
| Harness | Source | Confidence | Link | |||
|---|---|---|---|---|---|---|
| GPT-6 Astra | 69.4 | N/A | Independent | medium | Oct 15, 2026 | arxiv.org |
| Gemini 3.8 Flash | 68.4 | N/A | Independent | medium | Oct 15, 2026 | arxiv.org |
| GPT-5.6 Sol | 64.9 | N/A | Independent | medium | Oct 15, 2026 | arxiv.org |
| google/gemma-4-31B-it | 62.6 | N/A | Independent | medium | Oct 15, 2026 | arxiv.org |
| Qwen3.5 122B A10B | 61.6 | N/A | Independent | medium | Oct 15, 2026 | arxiv.org |
| Qwen3.6 35B A3B | 59.5 | N/A | Independent | medium | Oct 15, 2026 | arxiv.org |
| InternVL3.5-38B | 56.7 | N/A | Independent | medium | Oct 15, 2026 | arxiv.org |
| GLM-4.6V-106B | 55.0 | N/A | Independent | medium | Oct 15, 2026 | arxiv.org |
| OmniVerifier-7B | 52.0 | N/A | Independent | medium | Oct 15, 2026 | arxiv.org |
| Qwen2.5-VL-7B | 50.2 | N/A | Independent | medium | Oct 15, 2026 | arxiv.org |
Every number links to where it was published. Low-confidence rows are muted and need per-cell verification.
3DVerifierBench
Grounding Generative Verification in 3D ContentShuanglin Li, Jiaying Lin, Shuquan Ye. Paper: arXiv:2610.22360 The test set contains 2,581 samples across ten tasks, with 970 True and 1,611 False answers. It includes the original image/video inputs and available 3D assets. Related samples are grouped into 1,167 cases.
Test annotations
metadata/benchmark.jsonl contains one sample per line, with eight fields:
Field Meaning
sample_id Short…