Pelican Bicycle ArchiveAN AI FIELD GUIDE
中文EnglishРусскийPortuguês (Brasil)
THE COLLECTION

The collection · Page 30

1832 cases · Page 30 / 77

gpt-6-astra · official-direct · Sep 14 11:57 SVG previewImageCommunity record

gpt-6-astra · official-direct · Sep 14 11:57 SVG preview

One independent output in the 11:57 column of Yu Zhang's published SVG-code run board, labelled gpt-6-astra / official-direct. The author dates the runs to 2026-09-14. Row and column labels establish the displayed mapping. This is a faithful pixel crop of the original board; the full image is retained. Raw SVG responses are not published. Reasoning effort, actual provider identity and exact generation seconds are not independently verified. No timeline representative or channel-integrity verdict is inferred.

#843
gpt-6-astra
Nex-N2.5 Pro · 3D Bicycle Ride in Blender3D workCommunity record

Nex-N2.5 Pro · 3D Bicycle Ride in Blender

The author reports using Nex-N2.5 Pro through Pi and Blender MCP to generate an editable pelican bicycle scene, then requested an animation. The complete final animation is shown first; the original render and both full process recordings are also preserved. The recording shows code for the scene. The .blend project was not obtained.

#1815
Nex-N2.5 Pro
astra · xhigh · Three.jsImageCommunity record

astra · xhigh · Three.js

The date is the publication date shown for this source post / reply, not an authenticated generation date. Model and author labels follow the source and are not independently authenticated. Original preview bytes and provenance are preserved; views, frames and export formats of one project do not count as additional works. No explicit code licence was verified, so executable HTML is not mirrored and this is not an on-site Play demo. The reply only labels this astra xhigh; no product version or rendering engine is inferred from one preview.

#842
astra
Qwen-3.8-Max · procedural Three.js pelican ride3D workCommunity record

Qwen-3.8-Max · procedural Three.js pelican ride

Mapped to the corresponding output in the author’s eight-model comparison. The original crop and full comparison remain archived. The complete original Three.js animation from the pinned commit is restored with its Apache 2.0 license. Model and date are source-reported; restoring playback adds no work or Play count.

#841
Qwen-3.8-Max
DeepSeek-V4.1-Flash · procedural Three.js pelican ride3D workCommunity record

DeepSeek-V4.1-Flash · procedural Three.js pelican ride

Mapped to the corresponding output in the author’s eight-model comparison. The original crop and full comparison remain archived. The complete original Three.js animation from the pinned commit is restored with its Apache 2.0 license. Model and date are source-reported; restoring playback adds no work or Play count.

#840
DeepSeek-V4.1-Flash
Gemini-3.1-Pro-Preview · procedural Three.js pelican ride3D workCommunity record

Gemini-3.1-Pro-Preview · procedural Three.js pelican ride

Mapped to the corresponding output in the author’s eight-model comparison. The original crop and full comparison remain archived. The complete original Three.js animation from the pinned commit is restored with its Apache 2.0 license. Model and date are source-reported; restoring playback adds no work or Play count.

#839
Gemini-3.1-Pro-Preview
Claude-Sonnet-5 · procedural Three.js pelican ride3D workCommunity record

Claude-Sonnet-5 · procedural Three.js pelican ride

Mapped to the corresponding output in the author’s eight-model comparison. The original crop and full comparison remain archived. The complete original Three.js animation from the pinned commit is restored with its Apache 2.0 license. Model and date are source-reported; restoring playback adds no work or Play count.

#838
Claude-Sonnet-5
GLM-5.3 · procedural Three.js pelican ride3D workCommunity record

GLM-5.3 · procedural Three.js pelican ride

Mapped to the corresponding output in the author’s eight-model comparison. The original crop and full comparison remain archived. The complete original Three.js animation from the pinned commit is restored with its Apache 2.0 license. Model and date are source-reported; restoring playback adds no work or Play count.

#837
GLM-5.3
Kimi-K3 · procedural Three.js pelican ride3D workCommunity record

Kimi-K3 · procedural Three.js pelican ride

Mapped to the corresponding output in the author’s eight-model comparison. The original crop and full comparison remain archived. The complete original Three.js animation from the pinned commit is restored with its Apache 2.0 license. Model and date are source-reported; restoring playback adds no work or Play count.

#836
Kimi-K3
GPT-5.6 Sol · procedural Three.js pelican ride3D workCommunity record

GPT-5.6 Sol · procedural Three.js pelican ride

Mapped to the corresponding output in the author’s eight-model comparison. The original crop and full comparison remain archived. The complete original Three.js animation from the pinned commit is restored with its Apache 2.0 license. Model and date are source-reported; restoring playback adds no work or Play count.

#835
GPT-5.6 Sol
GPT-6-Astra · procedural Three.js pelican ride3D workCommunity record

GPT-6-Astra · procedural Three.js pelican ride

Mapped to the corresponding output in the author’s eight-model comparison. The original crop and full comparison remain archived. The complete original Three.js animation from the pinned commit is restored with its Apache 2.0 license. Model and date are source-reported; restoring playback adds no work or Play count.

#834
GPT-6-Astra
GPT-5.6 Sol · Wheels drift away from the frameAnimationCommunity record

GPT-5.6 Sol · Wheels drift away from the frame

The author posted the rendered result and the complete SVG in the same reply. The original ChatGPT screenshot shows the model label 5.6 Sol. The SVG rotates the wheel group as a unit, separating the wheels from the frame. Browser playback preserves this behavior.

#1679
GPT-5.6 Sol
Qwen3.8-flash-next · independent animation 9_x6mAnimationCommunity record

Qwen3.8-flash-next · independent animation 9_x6m

Independent source run 9_x6m, listed under xhigh. The original animated WebP is retained without re-encoding or repair. The author reports DGX Spark GB10, pi agent and cc-switch, and explicitly suspects an inverted low/medium mapping. Setting names are source labels, not authenticated backend levels. All 29 outputs count as works; multiple medium runs have no documented default representative, so none is guessed for the timeline. The date is publication, not an authenticated generation date.

#833
Qwen3.8-flash-next
Qwen3.8-flash-next · independent animation 8_x33mAnimationCommunity record

Qwen3.8-flash-next · independent animation 8_x33m

Independent source run 8_x33m, listed under xhigh. The original animated WebP is retained without re-encoding or repair. The author reports DGX Spark GB10, pi agent and cc-switch, and explicitly suspects an inverted low/medium mapping. Setting names are source labels, not authenticated backend levels. All 29 outputs count as works; multiple medium runs have no documented default representative, so none is guessed for the timeline. The date is publication, not an authenticated generation date.

#832
Qwen3.8-flash-next
Qwen3.8-flash-next · independent animation 8_mAnimationCommunity record

Qwen3.8-flash-next · independent animation 8_m

Independent source run 8_m, listed under medium. The original animated WebP is retained without re-encoding or repair. The author reports DGX Spark GB10, pi agent and cc-switch, and explicitly suspects an inverted low/medium mapping. Setting names are source labels, not authenticated backend levels. All 29 outputs count as works; multiple medium runs have no documented default representative, so none is guessed for the timeline. The date is publication, not an authenticated generation date.

#831
Qwen3.8-flash-next
Qwen3.8-flash-next · independent animation 8_l3mAnimationCommunity record

Qwen3.8-flash-next · independent animation 8_l3m

Independent source run 8_l3m, listed under low. The original animated WebP is retained without re-encoding or repair. The author reports DGX Spark GB10, pi agent and cc-switch, and explicitly suspects an inverted low/medium mapping. Setting names are source labels, not authenticated backend levels. All 29 outputs count as works; multiple medium runs have no documented default representative, so none is guessed for the timeline. The date is publication, not an authenticated generation date.

#830
Qwen3.8-flash-next
Qwen3.8-flash-next · independent animation 7_x16mAnimationCommunity record

Qwen3.8-flash-next · independent animation 7_x16m

Independent source run 7_x16m, listed under xhigh. The original animated WebP is retained without re-encoding or repair. The author reports DGX Spark GB10, pi agent and cc-switch, and explicitly suspects an inverted low/medium mapping. Setting names are source labels, not authenticated backend levels. All 29 outputs count as works; multiple medium runs have no documented default representative, so none is guessed for the timeline. The date is publication, not an authenticated generation date.

#829
Qwen3.8-flash-next
Qwen3.8-flash-next · independent animation 7_mAnimationCommunity record

Qwen3.8-flash-next · independent animation 7_m

Independent source run 7_m, listed under medium. The original animated WebP is retained without re-encoding or repair. The author reports DGX Spark GB10, pi agent and cc-switch, and explicitly suspects an inverted low/medium mapping. Setting names are source labels, not authenticated backend levels. All 29 outputs count as works; multiple medium runs have no documented default representative, so none is guessed for the timeline. The date is publication, not an authenticated generation date.

#828
Qwen3.8-flash-next
Qwen3.8-flash-next · independent animation 7_l2mAnimationCommunity record

Qwen3.8-flash-next · independent animation 7_l2m

Independent source run 7_l2m, listed under low. The original animated WebP is retained without re-encoding or repair. The author reports DGX Spark GB10, pi agent and cc-switch, and explicitly suspects an inverted low/medium mapping. Setting names are source labels, not authenticated backend levels. All 29 outputs count as works; multiple medium runs have no documented default representative, so none is guessed for the timeline. The date is publication, not an authenticated generation date.

#827
Qwen3.8-flash-next
Qwen3.8-flash-next · independent animation 6_mAnimationCommunity record

Qwen3.8-flash-next · independent animation 6_m

Independent source run 6_m, listed under medium. The original animated WebP is retained without re-encoding or repair. The author reports DGX Spark GB10, pi agent and cc-switch, and explicitly suspects an inverted low/medium mapping. Setting names are source labels, not authenticated backend levels. All 29 outputs count as works; multiple medium runs have no documented default representative, so none is guessed for the timeline. The date is publication, not an authenticated generation date.

#826
Qwen3.8-flash-next
Qwen3.8-flash-next · independent animation 6_l2mAnimationCommunity record

Qwen3.8-flash-next · independent animation 6_l2m

Independent source run 6_l2m, listed under low. The original animated WebP is retained without re-encoding or repair. The author reports DGX Spark GB10, pi agent and cc-switch, and explicitly suspects an inverted low/medium mapping. Setting names are source labels, not authenticated backend levels. All 29 outputs count as works; multiple medium runs have no documented default representative, so none is guessed for the timeline. The date is publication, not an authenticated generation date.

#825
Qwen3.8-flash-next
Qwen3.8-flash-next · independent animation 5_x17mAnimationCommunity record

Qwen3.8-flash-next · independent animation 5_x17m

Independent source run 5_x17m, listed under xhigh. The original animated WebP is retained without re-encoding or repair. The author reports DGX Spark GB10, pi agent and cc-switch, and explicitly suspects an inverted low/medium mapping. Setting names are source labels, not authenticated backend levels. All 29 outputs count as works; multiple medium runs have no documented default representative, so none is guessed for the timeline. The date is publication, not an authenticated generation date.

#824
Qwen3.8-flash-next
Qwen3.8-flash-next · independent animation 5_mAnimationCommunity record

Qwen3.8-flash-next · independent animation 5_m

Independent source run 5_m, listed under medium. The original animated WebP is retained without re-encoding or repair. The author reports DGX Spark GB10, pi agent and cc-switch, and explicitly suspects an inverted low/medium mapping. Setting names are source labels, not authenticated backend levels. All 29 outputs count as works; multiple medium runs have no documented default representative, so none is guessed for the timeline. The date is publication, not an authenticated generation date.

#823
Qwen3.8-flash-next