SupaVoxel
An AI image-to-3D model tool that supports exporting both compressed and uncompressed GLB files and performs notably in geometric data compression and engineering efficiency.
Timeline (last 90 days)
In the author's independent tests, Meshy scored 59/100 and SupaVoxel scored 89/100, with the scoring difference coming mainly from engineering efficiency rather than the models' visual appearance quality.
The graduation-monkey GLB file generated by Meshy is 69,179,580 bytes, while SupaVoxel's is 11,165,408 bytes; both contain 3 color textures each, and the Meshy file is about 6 times the size of SupaVoxel's.
On a 12 Mbps mobile network, downloading the 69.18 MB Meshy file takes about 46 seconds, while the 11.17 MB SupaVoxel file takes about 7 seconds.
The Meshy model used 2,086,002 triangles and SupaVoxel used 1,499,906 triangles, with Meshy having 586,096 more triangles.
SupaVoxel's face model has a subtle seam at the boundary between the bridge of the nose and the cheek, while Meshy has no visible seam.
Meshy's GLB file has an empty extensions list and can be opened without any decoder; SupaVoxel's compressed file declares two extensions, meshopt and quantization.
SupaVoxel's export menu offers two options: a compressed GLB of 11,165,408 bytes and an uncompressed GLB of 82,060,632 bytes at original size.
The graduation-monkey files generated by both tools contain zero skeletons and zero animation clips, so they cannot directly blink, wave or speak.
Estimated at a standard layout of 32 bytes per corner and 4 bytes per triangle index, Meshy's geometry requires about 61.17 MB of memory and SupaVoxel's about 46.50 MB, a gap of about 14.67 MB.
Both tools did not blur the four right angles of the graduation cap in the top-down view, performing comparably.
14 more timeline events
All Facts (20)
In this test, the Meshy exported model had 424,152 corners, while SupaVoxel had 865,508.
50%UnverifiedThis test did not record failures or retry counts, and Meshy's generation time was not captured, so there is no speed comparison data.
50%UnverifiedAfter cleanup, both the Meshy and SupaVoxel models are fully closed, with zero open edges, no non-manifold edges, consistent surface orientation, and each being a single connected shell.
50%UnverifiedBefore cleanup, Meshy reported 67,946 loose edges and SupaVoxel reported 225,446, all of which disappeared after merging vertices at the same positions.
50%UnverifiedIn this test, the Meshy exported file size was 27,408,396 bytes, while SupaVoxel's was 9,044,340 bytes.
50%UnverifiedAt a height of 120mm, the Meshy model's average edge length was 0.306mm, while SupaVoxel's was 0.295mm.
50%UnverifiedIn this image-to-3D comparison on the same Voronoi eggshell image, Meshy scored 61 and SupaVoxel scored 90; the scores are subjective and apply only to this task.
50%UnverifiedIn this test, Meshy's color image was 1,658,610 bytes and SupaVoxel's was 519,966 bytes, with Meshy having about three times the color data.
50%UnverifiedBoth Meshy and SupaVoxel can export in GLB format, with geometry and images packaged in the same file.
50%UnverifiedIn this test, the Meshy model had 419 stretched slender triangles and SupaVoxel had 2,849; Meshy had no degenerate near-zero-area faces, while SupaVoxel had two.
50%UnverifiedMeshy's GLB declares no required extensions, while SupaVoxel's compressed GLB declares EXT_meshopt_compression and KHR_mesh_quantization, requiring a decoder.
50%UnverifiedIn this test, the Meshy exported model had 779,020 triangles, while SupaVoxel had 1,500,000.
50%UnverifiedThe graduation-monkey files generated by both tools contain zero skeletons and zero animation clips, so they cannot directly blink, wave or speak.
50%UnverifiedThe graduation-monkey GLB file generated by Meshy is 69,179,580 bytes, while SupaVoxel's is 11,165,408 bytes; both contain 3 color textures each, and the Meshy file is about 6 times the size of SupaVoxel's.
50%UnverifiedOn a 12 Mbps mobile network, downloading the 69.18 MB Meshy file takes about 46 seconds, while the 11.17 MB SupaVoxel file takes about 7 seconds.
50%UnverifiedThe Meshy model used 2,086,002 triangles and SupaVoxel used 1,499,906 triangles, with Meshy having 586,096 more triangles.
50%UnverifiedSupaVoxel's face model has a subtle seam at the boundary between the bridge of the nose and the cheek, while Meshy has no visible seam.
50%UnverifiedMeshy's GLB file has an empty extensions list and can be opened without any decoder; SupaVoxel's compressed file declares two extensions, meshopt and quantization.
50%UnverifiedSupaVoxel's export menu offers two options: a compressed GLB of 11,165,408 bytes and an uncompressed GLB of 82,060,632 bytes at original size.
50%UnverifiedIn the author's independent tests, Meshy scored 59/100 and SupaVoxel scored 89/100, with the scoring difference coming mainly from engineering efficiency rather than the models' visual appearance quality.
50%