Benchmark Almanac

3D Arena

PicGen3D 1,429 tops its leaderboard
Not in model cards

Users compare two 3D models generated from the same input and vote for the better one. A live leaderboard ranks the generators.

The space was created in 2023-10 and was still updated on October 26th, 2026. It has about 415 likes.

Kind
Spatial reasoning and counting
Level
Everyday
Scored
Ranked by votes
Answers
Open-ended
Tasks
All public
28
Entries on its own leaderboard, 0 models we track

PicGen3D tops the leaderboard with 1,429

Best score per entry from the benchmark's own leaderboard, read on November 12th, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders.

28 entries. Scroll sideways for the rest. Bars start at 900.

  1. 1,429 PicGen3D
  2. 1,418 TRELLIS-3DGS
  3. 1,369 Meshy-6
  4. 1,346 404_GEN
  5. 1,333 TRELLIS
  6. 1,327 Meshy-7
  7. 1,318 Zaohaowu3D
  8. 1,312 Hunyuan3D-2
  9. 1,309 Hi3D-1.0
  10. 1,304 Hunyuan3D-2.1
  11. 1,284 SAM-3D-Objects-3DGS
  12. 1,283 UniLat3D-3DGS
  13. 1,273 InstantMesh
  14. 1,273 Pixal3D
  15. 1,267 TripoSG
  16. 1,234 Unique3D
  17. 1,207 Hi3DGen
  18. 1,186 SF3D
  19. 1,183 MeshFormer
  20. 1,172 TRELLIS.2-4B
  21. 1,151 Real3D
  22. 1,146 SPAR3D
  23. 1,102 UniLat3D-Mesh
  24. 1,099 TripoSR
  25. 1,083 LGM
  26. 1,064 Meshy-5
  27. 996 IM-MA
  28. 954 3DTopia-XL

Bars start at 900.

  1. 1 PicGen3D 1,429
  2. 2 TRELLIS-3DGS 1,418
  3. 3 Meshy-6 1,369
  4. 4 404_GEN 1,346
  5. 5 TRELLIS 1,333
  6. 6 Meshy-7 1,327
  7. 7 Zaohaowu3D 1,318
  8. 8 Hunyuan3D-2 1,312
  9. 9 Hi3D-1.0 1,309
  10. 10 Hunyuan3D-2.1 1,304
  11. 11 SAM-3D-Objects-3DGS 1,284
  12. 12 UniLat3D-3DGS 1,283
  13. 13 InstantMesh 1,273
  14. 14 Pixal3D 1,273
  15. 15 TripoSG 1,267
  16. 16 Unique3D 1,234
  17. 17 Hi3DGen 1,207
  18. 18 SF3D 1,186
  19. 19 MeshFormer 1,183
  20. 20 TRELLIS.2-4B 1,172
  21. 21 Real3D 1,151
  22. 22 SPAR3D 1,146
  23. 23 UniLat3D-Mesh 1,102
  24. 24 TripoSR 1,099
  25. 25 LGM 1,083
  26. 26 Meshy-5 1,064
  27. 27 IM-MA 996
  28. 28 3DTopia-XL 954

Source: Official 3D Arena leaderboard

#EntryScoreSettingDateSource
1 PicGen3D 1,429 Official leaderboard
2 TRELLIS-3DGS 1,418 Official leaderboard
3 Meshy-6 1,369 Official leaderboard
4 404_GEN 1,346 Official leaderboard
5 TRELLIS 1,333 Official leaderboard
6 Meshy-7 1,327 Official leaderboard
7 Zaohaowu3D 1,318 Official leaderboard
8 Hunyuan3D-2 1,312 Official leaderboard
9 Hi3D-1.0 1,309 Official leaderboard
10 Hunyuan3D-2.1 1,304 Official leaderboard
11 SAM-3D-Objects-3DGS 1,284 Official leaderboard
12 UniLat3D-3DGS 1,283 Official leaderboard
13 InstantMesh 1,273 Official leaderboard
14 Pixal3D 1,273 Official leaderboard
15 TripoSG 1,267 Official leaderboard
16 Unique3D 1,234 Official leaderboard
17 Hi3DGen 1,207 Official leaderboard
18 SF3D 1,186 Official leaderboard
19 MeshFormer 1,183 Official leaderboard
20 TRELLIS.2-4B 1,172 Official leaderboard
21 Real3D 1,151 Official leaderboard
22 SPAR3D 1,146 Official leaderboard
23 UniLat3D-Mesh 1,102 Official leaderboard
24 TripoSR 1,099 Official leaderboard
25 LGM 1,083 Official leaderboard
26 Meshy-5 1,064 Official leaderboard
27 IM-MA 996 Official leaderboard
28 3DTopia-XL 954 Official leaderboard

No lab model card reports it yet

We read the headline results of 522 launch posts from 52 labs, and none of them list this benchmark.

More in Vision

All 281 in Vision
  • MMMU In use 11.5K college-level multimodal questions across 30 subjects. Multiple-choice and open accuracy. GPT-5.1 85.4%
  • MMMU-Pro In use Harder MMMU: college-level multimodal questions with more options and vision-only inputs to remove text shortcuts. GPT-5.6 Sol 84.6%
  • CharXiv In use Chart understanding on real arXiv figures. Reasoning questions require synthesizing information from complex charts. Claude Fable 5 93.5%
  • MATH-Vision (MathVision) In use 3,040 competition math problems with visual context. Answer accuracy. Qwen 3.8 Max 97.7%
  • MMStar In use Models answer 1,500 human-curated image questions that cannot be solved without the image, across 6 core capabilities. Scored by accuracy. Qwen 3.8 Max 85.9%
  • ChartQA In use Models answer human-written and generated questions about chart images that need visual and arithmetic reasoning. Scored by relaxed accuracy with a 5 percent numeric tolerance. Command A Vision 90.9%