attempt-1 · Claude Code requested adaptive · infrastructure_interrupted
Requests / complete responses: 14 / 12
Early EEVEE infrastructure interruption; original evidence retained.
LongCat-2.5-Preview drives local Claude Code across 15 Blender, 3D, SVG and knowledge-video tasks, with actual artifacts, failures, API token usage and editable source downloads.
Updated: 2026-10-27
Read the model review: access, pricing, text, code and SVG tests
TOKRACE original creative evaluation · 2026-10-27
Local Claude Code drove LongCat through 15 production tasks: six Blender/3D scenes, four knowledge videos, four static SVGs and one SVG animation. Seven sessions ended normally; eight reached request, time or conservative usage limits. Session completion is not an artifact pass rate. The four static images meet the principal explicit requirements, and two bookshelf variants pass independent dimensional checks. Complex 3D scenes are mostly partial deliveries; three knowledge films exist with documented issues or limits, the email film is incomplete, and the cycling pelican has detached feet.
Token totals cover received final usage. Failed or interrupted requests can consume additional quota; their final usage is unknown. These tasks do not establish a general model ranking.
LongCat-2.5-Preview · LongCat API · Claude Code 2.1.311
15 tasks: six Blender/3D scenes, four knowledge videos, four static SVGs and one SMIL animation. Local Claude Code uses isolated offline shell, pixel feedback and SVG rendering tools, with at most two concurrent cases. LongCat generates and revises all artwork code.
The configured LongCat API supplies generation. Totals sum final usage across all attempts, including cache fields and thinking, excluding separate configuration probes. Missing usage remains unknown. Claude Code USD estimates are not LongCat invoices. Fixed local Kokoro narration is infrastructure, not native model speech.
LongCat writes, previews and revises artwork; Codex orchestrates and independently checks outputs. A sandbox EEVEE fault led to Cycles CPU continuation, with earlier attempts preserved. This is not an identical-condition comparison with historical Kimi runs.
01 · Blender · Requested effort: Claude Code requested adaptive
Scene and preview generated; early renderer faults and conservative budgeting prevented the full delivery.

Partial reducer preview: end plates obscure the internal gears
Web display uses WebP encoding only; original PNGs remain in the local evidence archive.
Original SHA-256: 5121c72e2467665498df2e33619f4b3f2c28c4de4cdc9d128b2ccd8f0db9960a
Public SHA-256: 365dcc768ab9b75f42e00398397ce29deafe0df33a2b097f2d4672c6f6749d3e
These are qualitative AI observations, not independent human scores. Automated checks are listed separately.
Task source and rerun specification
Independent offline generation preserving task requirements; Claude Code tool loop and Cycles CPU differ from the source benchmark.
The linked creator or official example is a source. The output above is our independently generated run.
24/18/60-tooth planetary reducer and axial explosion. Source inputs remain archived locally and are identified by hashes.
Archived source prompt SHA-256: 4b2e0fcc930a262160e2804f8bfbeadc8dc98c1df6f80f8f2a8a64e5133104dd
以下为TOKRACE新增执行条件,并非原benchmark协议:本轮离线,仅本题公开输入,不访问公网或外部模型/资产库;本批冻结的模型与参数,60分钟/50次请求/150000输出token,2线程。使用本机Blender5.2.3、EEVEE。原题的Cycles、OptiX、HDRI下载、采样/分辨率要求改为本次统一协议,必须在README披露;不声称官方BlenderWorld分数或严格复现原渲染管线。完整保留几何结构、尺寸、数量、配合、材质区分及原始UV/LOD等资产需求,不能为填任务数量简化;缺项如实列出。 先分功能块落盘完整可运行源代码,然后实际后台工厂启动建模/渲染,不使用单次巨型响应。交付final.blend、完整源码/资源、1920×1080主图+两张1600×900不同机位+结构/中性灰模近景;另交6秒1280×720/24fps的turntable.mp4(144帧)。减速器/蒸汽机可用机构运动代转台,但必须明确是运动学,非受力仿真,保留静态检查。先真实预览、view_image检查,最多两轮有效视觉修正,程序报错另记;保留首个预览与失败。视频检查至少6时点与首尾。独立进程重开工程,交artifacts.json/README(原题差异、实际命令版本、未完成项)/真实尺寸与对象清单自检JSON。离线程序材料替代外部图片/HDRI时必须说明;不得用平面或贴图代关键实体。 本题新增可测规格:读取asset_budget.json,全部几何/齿数/名义配合/UV/LOD要求保留。两张不同机位至少一张为同轴爆炸图,灰模近景看清太阳—行星—内圈接触。未实际校核的制造公差只能称名义展示。若运动,保持理论运动学关系,非动力学。 CPU protocol: Cycles CPU, 2 threads. Full private source inputs remain archived locally; source links and hashes identify them.
This is the evaluation adaptation and a rerun guide, not a verbatim transcript of the full original request. Exact original inputs remain archived locally.
Original evaluation input SHA-256: 4fe0a986df9baf3fd9f6d8be5fe1ea19510f2cc1e9ae22136301cf9ab95a6a70
Requests / complete responses: 14 / 12
Early EEVEE infrastructure interruption; original evidence retained.
Requests / complete responses: 15 / 15
Cycles CPU continuation stopped under conservative reservations for prior unknown usage; partial artifacts retained.
02 · Blender · Requested effort: Claude Code requested adaptive
A reopenable scene and partial preview exist; the complete still set and film were not delivered.

Partial steam-engine preview: pipes and metal surfaces; full delivery incomplete
Web display uses WebP encoding only; original PNGs remain in the local evidence archive.
Original SHA-256: 95af6b3809b0418eb53f8abe894fc79f3576338b26386bef803f5262cc97f810
Public SHA-256: 136969ba2195bdcc30aa76dd61cdf171e3f5f3be7913a348d1d8d216f1ff2d65
These are qualitative AI observations, not independent human scores. Automated checks are listed separately.
Task source and rerun specification
Independent offline generation preserving task requirements; Claude Code tool loop and Cycles CPU differ from the source benchmark.
The linked creator or official example is a source. The output above is our independently generated run.
Horizontal steam engine, flywheel and Watt governor. Source inputs remain archived locally and are identified by hashes.
Archived source prompt SHA-256: 46f733406de20668bf8ab71825ef43e931d42ca8d62946befaae89be3ae97ff4
以下为TOKRACE新增执行条件,并非原benchmark协议:本轮离线,仅本题公开输入,不访问公网或外部模型/资产库;本批冻结的模型与参数,60分钟/50次请求/150000输出token,2线程。使用本机Blender5.2.3、EEVEE。原题的Cycles、OptiX、HDRI下载、采样/分辨率要求改为本次统一协议,必须在README披露;不声称官方BlenderWorld分数或严格复现原渲染管线。完整保留几何结构、尺寸、数量、配合、材质区分及原始UV/LOD等资产需求,不能为填任务数量简化;缺项如实列出。 先分功能块落盘完整可运行源代码,然后实际后台工厂启动建模/渲染,不使用单次巨型响应。交付final.blend、完整源码/资源、1920×1080主图+两张1600×900不同机位+结构/中性灰模近景;另交6秒1280×720/24fps的turntable.mp4(144帧)。减速器/蒸汽机可用机构运动代转台,但必须明确是运动学,非受力仿真,保留静态检查。先真实预览、view_image检查,最多两轮有效视觉修正,程序报错另记;保留首个预览与失败。视频检查至少6时点与首尾。独立进程重开工程,交artifacts.json/README(原题差异、实际命令版本、未完成项)/真实尺寸与对象清单自检JSON。离线程序材料替代外部图片/HDRI时必须说明;不得用平面或贴图代关键实体。 本题新增可测规格:完整读取brief.md,底座、缸径行程、600mm连杆、1200mm飞轮的6条曲辐、离心调速器、滑阀机构和管路等原规格不缩水。4K材质贴图、UV、SM_命名、hero/LOD1/LOD2预算与运动副轴心需求保留;未完成或未烘焙须公开。不得只用彩色圆柱替代机械关系。增加背面、另一侧与中性灰模近景,观察汽缸—活塞杆—十字头—连杆—曲柄连接。视频可为转台或运动学机构;后者保持300mm行程、150mm曲柄、600mm杆长,不称蒸汽或受力仿真。 CPU protocol: Cycles CPU, 2 threads. Full private source inputs remain archived locally; source links and hashes identify them.
This is the evaluation adaptation and a rerun guide, not a verbatim transcript of the full original request. Exact original inputs remain archived locally.
Original evaluation input SHA-256: 408ff3a68c0d6d7c7011009f947598bc3b267d2ae19c01862c5a97d49ef3b58e
Requests / complete responses: 50 / 49
Early EEVEE infrastructure interruption; original evidence retained.
Requests / complete responses: 7 / 7
Cycles CPU continuation stopped under conservative reservations for prior unknown usage; partial artifacts retained.
03 · Blender · Requested effort: Claude Code requested adaptive
Scene and five final stills delivered; the session reached 50 requests without a final film.

Living-room view
Web display uses WebP encoding only; original PNGs remain in the local evidence archive.
Original SHA-256: 4d3a6984d40c13253d40354467b92199f4141ecc7bb09b4964b719ef4b18f3d2
Public SHA-256: c60e2fe98a1dfa176f496c40e0c07ee36b15d299b1aadfa247307205c2ae7323

Kitchen close-up
Web display uses WebP encoding only; original PNGs remain in the local evidence archive.
Original SHA-256: 895c30555ead1dbf9d64b0e70787da2bccf3740b682d3edcaeedde6fdf7e77ec
Public SHA-256: f2cc83b93f43fdc8c077a3bf1ae6c63e3d6822f18e284b83bbeb125d4f5aa7cd

Main bedroom
Web display uses WebP encoding only; original PNGs remain in the local evidence archive.
Original SHA-256: 668efc6f3c6e513350e6cbc5fbcc7c8a9f57b5f06bd801aa18c37e06c48a5076
Public SHA-256: 53eb0014348e136ef46e2025c71df4f5d2205dcdb13118c62c05e58211d9baf2

Apartment top view
Web display uses WebP encoding only; original PNGs remain in the local evidence archive.
Original SHA-256: f29345131d920ca1d8ca6d8bd80f5f6023cb3179157732f1cb889937f4cf799d
Public SHA-256: 0a93dbe31faaba6d244caa8cdd8ade1ceaf9fbe10e619af15ca4d4c99b22c0fe

Clay close-up: camera close to a wall
Web display uses WebP encoding only; original PNGs remain in the local evidence archive.
Original SHA-256: 0577df26b96c7e303654694a249e08ade9f8fc9ba31201cc5274f915585dcb25
Public SHA-256: f12afc9d87e133d4b473d6796a1e430dc05fa1a88e4526c7c0dd5830dd4e0540
These are qualitative AI observations, not independent human scores. Automated checks are listed separately.
Task source and rerun specification
Independent offline generation preserving task requirements; Claude Code tool loop and Cycles CPU differ from the source benchmark.
The linked creator or official example is a source. The output above is our independently generated run.
12×9m two-bedroom interior, rooms and floor plan. Source inputs remain archived locally and are identified by hashes.
Archived source prompt SHA-256: 3947abd2d7e39da19d44da55cc45bb9ceeba7220bd9d25b9ba44485f6f3c8579
以下为TOKRACE新增执行条件,并非原benchmark协议:本轮离线,仅本题公开输入,不访问公网或外部模型/资产库;本批冻结的模型与参数,60分钟/50次请求/150000输出token,2线程。使用本机Blender5.2.3、EEVEE。原题的Cycles、OptiX、HDRI下载、采样/分辨率要求改为本次统一协议,必须在README披露;不声称官方BlenderWorld分数或严格复现原渲染管线。完整保留几何结构、尺寸、数量、配合、材质区分及原始UV/LOD等资产需求,不能为填任务数量简化;缺项如实列出。 先分功能块落盘完整可运行源代码,然后实际后台工厂启动建模/渲染,不使用单次巨型响应。交付final.blend、完整源码/资源、1920×1080主图+两张1600×900不同机位+结构/中性灰模近景;另交6秒1280×720/24fps的turntable.mp4(144帧)。减速器/蒸汽机可用机构运动代转台,但必须明确是运动学,非受力仿真,保留静态检查。先真实预览、view_image检查,最多两轮有效视觉修正,程序报错另记;保留首个预览与失败。视频检查至少6时点与首尾。独立进程重开工程,交artifacts.json/README(原题差异、实际命令版本、未完成项)/真实尺寸与对象清单自检JSON。离线程序材料替代外部图片/HDRI时必须说明;不得用平面或贴图代关键实体。 本题新增可测规格:完整读取layout.yaml。户型12×9m、净高2.8m、外墙240mm和隔墙120mm、家具空间要求保持。增加三张室内机位、一张正交顶视和一张门窗/隔墙或家具尺度的灰模近景,因此本题至少5张最终静帧。原题4K/OptiX/Cycles改为本次1920主图、1600副图、EEVEE是公开执行差异。按房间分集合,交可核对墙界、门洞、家具数量和尺寸的JSON。南向大窗与附件通用窗台规格若有解释冲突,说明选择,不能假称唯一真值。 CPU protocol: Cycles CPU, 2 threads. Full private source inputs remain archived locally; source links and hashes identify them.
This is the evaluation adaptation and a rerun guide, not a verbatim transcript of the full original request. Exact original inputs remain archived locally.
Original evaluation input SHA-256: 0fb2e735ddb6f50f6ce3688cd537d6ed5214e64ab5529147d158520182297114
Requests / complete responses: 50 / 50
CLI reached 50 requests. Delivered artifacts are assessed independently; the exit code is not a quality score.
04 · Blender · Requested effort: Claude Code requested adaptive
Scene, four final stills, a six-second orbit film and documentation were delivered; the CLI still exited at its 50-request limit.
Download MP4 · Playback starts only when you press play.
Six-second courtyard orbit with fixed Kokoro Chinese narration and optional subtitles.

Main hall viewed from the entrance
Web display uses WebP encoding only; original PNGs remain in the local evidence archive.
Original SHA-256: 58182ded4f558d494a0d759c74f46bc2e8e468339225f4343b6d8f8cb0e9d6ed
Public SHA-256: d3c72464249d2d06519b5428f247f0b499e443c13dfd2d18ebf83008cfdc9415

Side view of the courtyard
Web display uses WebP encoding only; original PNGs remain in the local evidence archive.
Original SHA-256: bde33dc6fecf6062e9d1d64fac302423d3dffe5469f3808215ef15c704962c30
Public SHA-256: 826bf7fe1448400f89a6834502b61f6c1cfd4b5941f9b6fd89f079024139fef6

Top view with large black side margins
Web display uses WebP encoding only; original PNGs remain in the local evidence archive.
Original SHA-256: 98e1ac4997dd7e33054f87f01b006664a1f1c4eb5f5633e433e3c60385dcf999
Public SHA-256: 7af2276f86befade71e6117fb6ed2b4cd4d1d33e78457156f257fbdfe62dee50

Clay close-up of the eaves and beams
Web display uses WebP encoding only; original PNGs remain in the local evidence archive.
Original SHA-256: 181fb957fc08c117f4bb19aaffeb21ac35a970ef847343ad63b5d3feac9eed34
Public SHA-256: d2e1d9dd9096ac95490456784264db0f5c9f088f3ec845ab81fb756dd62b491e
These are qualitative AI observations, not independent human scores. Automated checks are listed separately.
Task source and rerun specification
Independent offline generation preserving task requirements; Claude Code tool loop and Cycles CPU differ from the source benchmark.
The linked creator or official example is a source. The output above is our independently generated run.
Traditional Chinese temple courtyard and roof structure. Source inputs remain archived locally and are identified by hashes.
Archived source prompt SHA-256: a21d6b051543e08ce7b43405090a6ca4f7fd1e1f46f4a4038036f66114c43e11
以下为TOKRACE新增执行条件,并非原benchmark协议:本轮离线,仅本题公开输入,不访问公网或外部模型/资产库;本批冻结的模型与参数,60分钟/50次请求/150000输出token,2线程。使用本机Blender5.2.3、EEVEE。原题的Cycles、OptiX、HDRI下载、采样/分辨率要求改为本次统一协议,必须在README披露;不声称官方BlenderWorld分数或严格复现原渲染管线。完整保留几何结构、尺寸、数量、配合、材质区分及原始UV/LOD等资产需求,不能为填任务数量简化;缺项如实列出。 先分功能块落盘完整可运行源代码,然后实际后台工厂启动建模/渲染,不使用单次巨型响应。交付final.blend、完整源码/资源、1920×1080主图+两张1600×900不同机位+结构/中性灰模近景;另交6秒1280×720/24fps的turntable.mp4(144帧)。减速器/蒸汽机可用机构运动代转台,但必须明确是运动学,非受力仿真,保留静态检查。先真实预览、view_image检查,最多两轮有效视觉修正,程序报错另记;保留首个预览与失败。视频检查至少6时点与首尾。独立进程重开工程,交artifacts.json/README(原题差异、实际命令版本、未完成项)/真实尺寸与对象清单自检JSON。离线程序材料替代外部图片/HDRI时必须说明;不得用平面或贴图代关键实体。 本题新增可测规格:本题增加对侧、俯视整体与屋顶/柱梁灰模近景。原题未固定长度,按合理古建尺度自主设计并记录,自定尺度不写成原题。石狮为一对实体,红柱/金顶/上翘屋檐、青石直达路径、铜钟与红灯笼应清楚可辨。 CPU protocol: Cycles CPU, 2 threads. Full private source inputs remain archived locally; source links and hashes identify them.
This is the evaluation adaptation and a rerun guide, not a verbatim transcript of the full original request. Exact original inputs remain archived locally.
Original evaluation input SHA-256: 038b0b7a257d216edda4c93d4c9e6ce06f1a210cd4d3af500616617444ae3435
Requests / complete responses: 50 / 50
CLI reached 50 requests. Delivered artifacts are assessed independently; the exit code is not a quality score.
05 · Blender · Requested effort: Claude Code requested adaptive
A reopenable drydock scene and four previews exist; final stills, film and documentation were not completed within 50 requests.

Main preview of the ship and maintenance deck
Web display uses WebP encoding only; original PNGs remain in the local evidence archive.
Original SHA-256: a4e3630784cc2ddefd067c1bc3bf1036a6260f4dc67b854d7606935587d85574
Public SHA-256: c30bd9f4dd021e2ea59079fd31cbe7b03f814595baa2d2da6adfc531eb90a889

Wide drydock preview
Web display uses WebP encoding only; original PNGs remain in the local evidence archive.
Original SHA-256: ac6190c417065a6023af9f168ec91f975b7a4a778a96d9692d2c9ce31e04faa6
Public SHA-256: 6163da64d999fea8d88deee1ddcec40533f842a1b58259f0ebab6e7621adb5e5

Stern and engine-opening preview
Web display uses WebP encoding only; original PNGs remain in the local evidence archive.
Original SHA-256: c3dbeef3e4ee62efa891027c337c837a67cae16f91c26115cd3ebf57768e848c
Public SHA-256: de1e67b1bdba728458b79ef46cb8cdd6ec42427c183b3419c68af211fd8bbf53

Hull close-up, with emissive materials still visible
Web display uses WebP encoding only; original PNGs remain in the local evidence archive.
Original SHA-256: d6f7c1d703d40bac2f5e2d1807b427ac84e54968a5e9fd8b3776ab34e59d0df9
Public SHA-256: 687b5f51f0443fdcde58c8d9b4b70534222bf88c988ea7713d4500da6df6d980
These are qualitative AI observations, not independent human scores. Automated checks are listed separately.
Task source and rerun specification
Independent offline generation preserving task requirements; Claude Code tool loop and Cycles CPU differ from the source benchmark.
The linked creator or official example is a source. The output above is our independently generated run.
Jovian orbital starship repair dock. Source inputs remain archived locally and are identified by hashes.
Archived source prompt SHA-256: 93a8724df0253fe95a3864cd8b2b075435b69a499043237ab1bb15971c0ec730
本次来源复跑说明:仅运行本批指定的一款模型,不与原作者K3桌面产品结果混为同条件排名。原题中的旧机器路径与测试根目录已替换为当前路径,其他场景与交付要求保留;固定EEVEE与本次工具预算属于本次条件。任务根目录就是当前工作目录,允许其下work/和outputs/。静态主图、两张不同角度图、灰模近景和可编辑工程均需完成,不能以单张概念图代替。 CPU protocol: Cycles CPU, 2 threads. Full private source inputs remain archived locally; source links and hashes identify them.
This is the evaluation adaptation and a rerun guide, not a verbatim transcript of the full original request. Exact original inputs remain archived locally.
Original evaluation input SHA-256: b6a4b6ed33b05f86761e07619bf0167adbe122d10e95c06d62245300dd69e1d9
Requests / complete responses: 50 / 48
CLI reached 50 requests. Delivered artifacts are assessed independently; the exit code is not a quality score.
06 · Blender · Requested effort: Claude Code requested adaptive
One parametric script produces both variants; independent geometry checks and multi-view images satisfy the fixed specifications.

Three-compartment baseline, 800mm wide
Web display uses WebP encoding only; original PNGs remain in the local evidence archive.
Original SHA-256: 7a87421eba2fdc70cf047bc7b8317826092ec9077de199b22f30b9a640e2d201
Public SHA-256: 06d039764c28360501e15c9db538aff482155283b75629577a1dc005354ad517

Baseline rear view with an open back
Web display uses WebP encoding only; original PNGs remain in the local evidence archive.
Original SHA-256: 73538435f2cf47b1a2da44bb0d456a1889c08b9ecd6c5b9993cafb21122f2ccf
Public SHA-256: 5ee40e3d29afd04b21eb5ed7670b367e8c657968392541c744d1cd17de557a6f

Four-compartment revision, 1000mm wide
Web display uses WebP encoding only; original PNGs remain in the local evidence archive.
Original SHA-256: c6164fe6a0e2a7066427bd33d0c4711d2bd3f5c1e83cb0a75d9ad3cb987cbb23
Public SHA-256: c939309ae9e3668bbbf7be7306ff9eadb79af57b14bcf655a02f4ca0fc8ae718

Baseline oblique view
Web display uses WebP encoding only; original PNGs remain in the local evidence archive.
Original SHA-256: 09d7ae1e845196ed61bbc8c718690abe87282772abdc3ac6ddd4da47b9c2a736
Public SHA-256: e593e43dd307679178b2042deb5dcb280b7c6446851ced94ac863f9eb9cb6441

Revised oblique view
Web display uses WebP encoding only; original PNGs remain in the local evidence archive.
Original SHA-256: d295fa036abe0ce08a8bdb99a7e94375332d6c4ab5eafa531db24e65f1c0cd7e
Public SHA-256: e2721389d03751e7813a8afa7bd33c146c5e6deb170a7fadbcb7c2e11811d1b9
These are qualitative AI observations, not independent human scores. Automated checks are listed separately.
Task source and rerun specification
Independent offline generation preserving task requirements; Claude Code tool loop and Cycles CPU differ from the source benchmark.
The linked creator or official example is a source. The output above is our independently generated run.
Parametric bookshelf and parameter revision. Source inputs remain archived locally and are identified by hashes.
Archived source prompt SHA-256: 86151fda84cd34fb1612185d4bf3edc051e49c38fd163d190b9acf2ae186f079
保留官方原题,增加下列可验收要求: 使用Blender原生几何制作可编辑木质书架,交付一个可通过命令行参数重建的Python脚本。 基础版外框宽800、深300、高1200毫米;全部板厚18毫米;由2块侧板、1块顶板、1块底板、2块内部水平隔板组成,形成3个等高净空储物格;无背板、无额外脚。顶底板嵌在两侧板之间,所有外边界尺寸必须一致。每块板独立命名,集中放在Bookcase集合;展示道具可有少量书,但不得混入Bookcase集合影响测量。 脚本公开width_mm、depth_mm、height_mm、thickness_mm、compartments、wood_color参数。用同一脚本再生成宽1000、高1200、深300、板厚18、4个等高储物格的修改版,不要另写独立几何代码。 交付baseline.blend和modified.blend、对应真实尺寸/对象清单JSON、至少3张1920×1080图片(基础版主视角、背面、修改版主视角),和清楚说明层数语义的README。模型背面要可检查,不能用背景平面代替实际建模。 先渲染低分辨率预览并用view_image检查。最多两轮有效视觉修正;执行报错可以另行修复。最终用独立Blender进程重开两个工程,确认集合、材质、相机和可测尺寸。保留脚本、预览、日志,不制作动画。 CPU protocol: Cycles CPU, 2 threads. Full private source inputs remain archived locally; source links and hashes identify them.
This is the evaluation adaptation and a rerun guide, not a verbatim transcript of the full original request. Exact original inputs remain archived locally.
Original evaluation input SHA-256: 84b4b32f2cbc33747cefca2f98ceb8003bdc0a71c79b266de9d5bdea6334b246
Requests / complete responses: 19 / 19
Model session finished normally; independent checks are separate.
07 · Knowledge video · Requested effort: Claude Code requested adaptive
An 89.84-second circulation film was generated with broadly consistent key diagrams and subtitles; the CLI reached 50 requests before completing every delivery manifest.
Download MP4 · Playback starts only when you press play.
Three-cell atmospheric circulation with fixed Kokoro Chinese narration and bilingual subtitles.

Actual film frame: 59.894s
Frame extracted from the original MP4 and encoded as WebP; content is not edited and original film bytes are retained.
Original SHA-256: e550646caa12abd8d6eb451037b173428f0c0af4fef27acf37606f6aeb7d06a9
Public SHA-256: 50b5e64a51dc8f25b32351e414f1ceb3102ce18a675f87ee5c119ccb21bd947c

Actual film frame: 69.876s
Frame extracted from the original MP4 and encoded as WebP; content is not edited and original film bytes are retained.
Original SHA-256: 095376c5530926350373bacaa349f6e821f640457079c1c505914ba6eef979a9
Public SHA-256: af39897b328d5e4234f3de24564aaa44514ca71bdaddefa5521ceee0ef3ca766
These are qualitative AI observations, not independent human scores. Automated checks are listed separately.
Task source and rerun specification
Independent offline generation preserving task requirements; Claude Code tool loop and Cycles CPU differ from the source benchmark.
The linked creator or official example is a source. The output above is our independently generated run.
Atmospheric circulation: three cells and deflected winds. Source inputs remain archived locally and are identified by hashes.
Archived source prompt SHA-256: b021ced84260e927afb5f3cc86e16892c9957aa363def7c92ee2f5b012d3e897
本次追加规格(与公开作者原prompt分开;以此消解原prompt中在线TTS配置要求) 任务:制作75–90秒中文大气环流知识短片,目标约90秒。保留线稿动画、中英双语字幕、合适的轻量程序化音乐,交付可直接导出的完整视频。 变化说明:原作者案例为Opus5.5、288.133秒1920×1080;本次使用本批指定模型,压缩为75–90秒,TTS换为本地固定Kokoro中文音色,因此TTS.md、.env/APIKEY/VOICE和云水印参数均不适用。不是原片逐镜头复刻。 画幅1920×1080,30fps,有中文配音、烧录中英字幕及SRT。配音只用本地固定Kokoro中文音色,不等待key;音乐若制作需程序化且旁白优先。 内容:解释不均匀加热和理想化三圈环流,重点画北半球0/约30/约60/90度的剖面。气团须沿每个闭合路径连续运动;对应地表信风、西风和极地东风,箭头与旁白一致。另用很短的地球/平面方向图解释北半球偏向运动方向的右侧,而不是‘永远向东’。注明南半球偏左以及真实气流比理想模型复杂、纬度边界会移动。以高中生能理解的方式讲一个实际含义,如副热带下沉与干燥。 不得把Ferrel环流说成简单独立热对流;不要求完整天气预报或喷流专题。纬度30/60为示意近似,不是固定自然边界。 执行只用已提供facts、Pillow/NumPy+FFmpeg、本地固定Kokoro中文音色;只读inputs,不联网或安装陌生代码。从空目录自行编写原创制片源码。 交付MP4、旁白WAV、SRT、讲稿、完整源码、重建说明、工具/素材来源、至少8张实际编码成片代表帧(包括三圈路径关键状态和结尾)。每次失败与重试要保留,不手工替模型修动画。 CPU protocol: Cycles CPU, 2 threads. Full private source inputs remain archived locally; source links and hashes identify them.
This is the evaluation adaptation and a rerun guide, not a verbatim transcript of the full original request. Exact original inputs remain archived locally.
Original evaluation input SHA-256: 9bcc49b2a402611500adfc2051ece2827dacd16062f8181d1b9e370bd8562fa3
Requests / complete responses: 50 / 50
CLI reached 50 requests. Film and README exist, but not every delivery manifest was completed.
08 · Knowledge video · Requested effort: Claude Code requested adaptive
An 81.9-second film and source were delivered. Core formulas are correct, with issues in the sign transition and local label readability.
Download MP4 · Playback starts only when you press play.
Secants approaching a tangent, with fixed Kokoro Chinese narration and Chinese subtitles.

Actual film frame: 18.2s
Frame extracted from the original MP4 and encoded as WebP; content is not edited and original film bytes are retained.
Original SHA-256: 741b3193d823a8a236a1bd8ebdf93aaed3a8da2b19f56830f9ae2299b2b04626
Public SHA-256: 7e7bf432f3f1613a86aac304234a39bb8407bbe92036530ffc50f4efc2fe6481

Display precision issue: 37.4s
Frame extracted from the original MP4 and encoded as WebP; content is not edited and original film bytes are retained.
Original SHA-256: 4cdf34711025423ecc53f362ccae22a68758d74bfb6011d35748a0bb3edb13f6
Public SHA-256: 8b81bd1260bb42ce367f5471753d7db99ab5830a98cf1ac40cc1ed377b3dfbd8
These are qualitative AI observations, not independent human scores. Automated checks are listed separately.
Task source and rerun specification
Independent offline generation preserving task requirements; Claude Code tool loop and Cycles CPU differ from the source benchmark.
The linked creator or official example is a source. The output above is our independently generated run.
Derivatives: a secant approaching the tangent. Source inputs remain archived locally and are identified by hashes.
Archived source prompt SHA-256: 5287dcf00a500b81f8363e9f8cfebeb93a90fd8f13440efe4ce0425bd0895383
本次追加规格(与公开作者原prompt分开) 制作75–90秒中文导数概念学习短片,目标约90秒。原作者为Opus5.5+Manim、edge-tts,公开视频457秒1920×1080;本次使用本批指定模型,只保留‘通俗易懂、有例子、引人思考’的主题要求,压缩成一个机制并用本地固定Kokoro中文音色替换edge-tts。不得称作完整原片复刻或Manim基准。 本轮不安装Manim/Remotion;使用Pillow、NumPy、FFmpeg离线原创动画。1920×1080、30fps、中文旁白、中文烧录字幕及SRT;轻量程序化音乐可选,旁白优先。 机制核心:以f(x)=x^2在P=(1,1)处为例。持续移动Q=(1+h,(1+h)^2),同步更新割线、斜率三角形、h值与差商数值。让正h逐步缩小,随后演示非零负h从另一侧靠近;最后得到斜率2、切线y=2x-1。不得在h=0直接除0,不把多个互不连续的静态卡片当作割线向切线的过程。 开头提出具体问题‘这一瞬间到底有多陡’,结尾让观众预测f(x)=x^2在x=2的斜率并给短暂停留,随后给答案或明确为思考题。图上刻度、位置、线斜率和数值必须来自同一数学状态;负h的水平与垂直变化要保持符号一致。 交付MP4、旁白WAV、SRT、讲稿、完整原创源码、可重建命令和环境说明、工具/来源说明,至少8张从实际编码MP4抽取的代表帧,覆盖h较大、趋近、负h、切线、结尾。每次失败/重试保留。 源事实和技巧已在inputs,执行不得联网或安装陌生代码,不等待key。 CPU protocol: Cycles CPU, 2 threads. Full private source inputs remain archived locally; source links and hashes identify them.
This is the evaluation adaptation and a rerun guide, not a verbatim transcript of the full original request. Exact original inputs remain archived locally.
Original evaluation input SHA-256: 8d3372c2200bbf52b2161241fb0d71b5211590524f606f51b11aa93ac3456a88
Requests / complete responses: 40 / 40
Model session finished normally. Independent review finds a display-precision issue rather than accepting every self-check claim.
09 · Knowledge video · Requested effort: Claude Code requested adaptive
A 117-second printer film was delivered, with an incorrect exposure beam and an interoperability problem in the original SRT timestamps.
Download MP4 · Playback starts only when you press play.
Original evaluation film, with an incorrect exposure diagram. Fixed Kokoro Chinese narration; a separate web VTT normalizes timestamps to the script intention.

13s frame: inkjet and laser family cards
Frame extracted from the original MP4 and encoded as WebP; content is not edited.
Original SHA-256: 8ff93afec4e116b587411602470e75650b709103832b9810a499e78775e24edf
Public SHA-256: f95fca7b6341519c0efdff06402a774ec5def5be6a35a10649721beb66d89f1c

78s frame: the beam passes through the photoconductor drum
Frame extracted from the original MP4 and encoded as WebP; content is not edited.
Original SHA-256: f4eb29d674eeb24c7900f0fd49410b9501db48a49a975a320fdaaf5a86a31ebd
Public SHA-256: 1d9d5418cd193ae88458ed75631beb193f6ea5fe7ec9174c0b87632dbd5b87a1
These are qualitative AI observations, not independent human scores. Automated checks are listed separately.
Task source and rerun specification
Independent offline generation preserving task requirements; Claude Code tool loop and Cycles CPU differ from the source benchmark.
The linked creator or official example is a source. The output above is our independently generated run.
A hand-drawn explainer of printer operation. Source inputs remain archived locally and are identified by hashes.
Archived source prompt SHA-256: ef251d6c101cd8017f2169ef06b68b5c3eff683b7eea98e1a858ebbc59d802be
本次新增可验收规格(原作者未披露时长/画幅/工具链,故不是原片逐镜头复制): 制作90至120秒、1920×1080、30fps的中文科普MP4,含清晰中文旁白与中文字幕/独立SRT。整体保持手绘线稿风格,至少5个具有因果解释的动态场景,不能仅是文字PPT。 面向普通读者,区分喷墨和激光两种技术:墨滴如何到纸上、激光如何在感光鼓上形成潜像、显影/转印/热压定影。范围和技术名称要明确;画面、运动方向、材料标签与讲稿一致。 已提供两份Canon官方事实材料到inputs/,可以阅读并在讲稿说明来源。不得把激光画成直接把文字烧在纸上。 使用统一条件中的固定Kokoro配音命令,不下载新TTS。所有画面由你写代码绘制和动画;不得调用原生图像/视频模型。交付video.mp4、narration.wav或m4a、captions.srt、script.txt、可完整重新生成的动画源码、至少8张带时间标记的代表帧、README和artifacts.json。 你要实际完成配音、渲染和混音,检查时长、分辨率、帧率、音轨存在及字幕起止。不要为了省时间缩短到不足90秒;不要把解说字幕当作全部视觉内容。 CPU protocol: Cycles CPU, 2 threads. Full private source inputs remain archived locally; source links and hashes identify them.
This is the evaluation adaptation and a rerun guide, not a verbatim transcript of the full original request. Exact original inputs remain archived locally.
Original evaluation input SHA-256: ad8a621c8c582fcf70027266562b5d2ac027abefaa50b39536aec1e3b165c61b
Requests / complete responses: 50 / 50
CLI ended at 50 requests. Film, source, self-checks and documentation exist; independent review does not accept every pass claim.
10 · Knowledge video · Requested effort: Claude Code requested adaptive
UI previews and generated source are available. The 60-minute limit ended the run without the required 75–85-second complete film.

Standalone preview from the incomplete project: Email-authentication title. Not a completed film.
Original preview converted to WebP without content correction.
Original SHA-256: bc1be9468c46221074dd2d937719b0daccb4ae15015782ab4c85a7006c0ec755
Public SHA-256: 2d87180275fadd12f8103fd44ca526480c6d69aa2bd026c3d0c5e445d8a642bb

Standalone preview from the incomplete project: SPF, DKIM and DMARC cards. Not a completed film.
Original preview converted to WebP without content correction.
Original SHA-256: fb2d92bec5cd39c09a9b4784355653ee4f05f389f209b86eb3f35de2a5c995bb
Public SHA-256: 86efac593454e44a217adfdb4580b6eb6885c79f2303e2f05200636701734b2a

Standalone preview from the incomplete project: Authentication versus content safety. Not a completed film.
Original preview converted to WebP without content correction.
Original SHA-256: f6ee6371d02da16cde7192b3aae9eaea47084cc310f8dff226be23f65d88db4d
Public SHA-256: efbd754181f638b930b2c1be4055e6f4705bca6a8e72255a7d938916dbd2f7df
These are qualitative AI observations, not independent human scores. Automated checks are listed separately.
Task source and rerun specification
Independent offline generation preserving task requirements; Claude Code tool loop and Cycles CPU differ from the source benchmark.
The linked creator or official example is a source. The output above is our independently generated run.
A motion-graphics guide to spoofed email. Source inputs remain archived locally and are identified by hashes.
Archived source prompt SHA-256: 1982b9439ae02d4828e0f5aca5cba72618cd25c6462778d7c046bdffd59a9398
Evaluation additions and limits: the source's earlier YouTube style context is unavailable. Create an original visual treatment, not a claimed exact style replica. The original media is about74.2s at1920×1080; this run requires75–85 seconds,1920×1080,30fps, English narration and synchronized English captions/SRT. Make a polished motion-graphics explainer for nontechnical viewers, with at least5 animated scenes that explain causality rather than just text slides. Use clearly fictional example.com/example.org addresses and illustrative messages; do not access or send real email. Read the supplied official Gmail authentication material and RFC9989 in inputs/ when needed. Explain visible sender identity, trusted original/header information and SPF/DKIM/DMARC at an accessible level; retain the limits of what authentication can prove. Do not imply a pass proves harmless content or a failure proves fraud. If you use Authentication-Results, show results added by the receiving service, not arbitrary untrusted text as authoritative. Use the fixed Kokoro English narration command in the shared runtime conditions. You must write all visual/animation code yourself and actually render/export. Deliver video.mp4, narration.wav or m4a, captions.srt, script.txt, complete rebuildable source, at least8 timestamped representative frames, README and artifacts.json. Confirm duration, dimensions,30fps, audio presence, captions and complete ending. Do not use image/video/music generation models or download large packages. The base renderer is your choice among installed tools. CPU protocol: Cycles CPU, 2 threads. Full private source inputs remain archived locally; source links and hashes identify them.
This is the evaluation adaptation and a rerun guide, not a verbatim transcript of the full original request. Exact original inputs remain archived locally.
Original evaluation input SHA-256: 3ef2665d13d52a3faacd9cdef32edf2c62e137de578fb8d0c97cbeb6b731fc43
Requests / complete responses: 44 / 44
Stopped at the 60-minute limit after 44 requests; no complete film.
11 · SVG · Requested effort: Claude Code requested adaptive
Recognizable pelican and bicycle, with a long bill, throat pouch, connected fork and frame.

Actual SVG render from this case
The model SVG is rasterized and encoded as WebP for display; source SVG is not rewritten.
Original SHA-256: ede5d8a31550b1494f87845f7103cfbf51a4e40bd9e3785ae1d9417a380aae89
Public SHA-256: 6bbcc3f74ffe1d69dd75f3be0a81ecd82f085b6131b0b56e257c58301b496467
These are qualitative AI observations, not independent human scores. Automated checks are listed separately.
Task source and rerun specification
Independent offline generation preserving task requirements; Claude Code tool loop and Cycles CPU differ from the source benchmark.
The linked creator or official example is a source. The output above is our independently generated run.
Pelican riding a bicycle. Source inputs remain archived locally and are identified by hashes.
Generate an SVG of a pelican riding a bicycle CPU protocol: Cycles CPU, 2 threads. Full private source inputs remain archived locally; source links and hashes identify them.
This is the evaluation adaptation and a rerun guide, not a verbatim transcript of the full original request. Exact original inputs remain archived locally.
Original evaluation input SHA-256: 2310a3a36c4b8999751f7562e91ef2991fb424faa7f9207aa414190c597436bf
Requests / complete responses: 14 / 14
Model session finished normally; first SVG and revisions are preserved, with semantic checks reported separately.
12 · SVG · Requested effort: Claude Code requested adaptive
The 09:41 clock passes independent SVG checks for numbers, hand angles and relative lengths.

Actual SVG render from this case
The model SVG is rasterized and encoded as WebP for display; source SVG is not rewritten.
Original SHA-256: 76058fb8a4ae8bba802965b315c27de143fd0241cf72f0cac6fa4d72aadd5b86
Public SHA-256: e40d23458a57012558e87926d2902ef5941bbb9d3325c9837067dd2f31a2a7c9
These are qualitative AI observations, not independent human scores. Automated checks are listed separately.
Task source and rerun specification
Independent offline generation preserving task requirements; Claude Code tool loop and Cycles CPU differ from the source benchmark.
The linked creator or official example is a source. The output above is our independently generated run.
An analog clock at 9:41. Source inputs remain archived locally and are identified by hashes.
Return only a complete, self-contained static SVG with viewBox="0 0 800 600". Draw a clear analog clock showing exactly 9:41, with all twelve numbers 1 through 12. The hour hand must be correctly positioned between 9 and 10, not exactly on 9. The minute hand must indicate minute 41. Use a shorter hour hand and a longer minute hand, no second hand. Add a small label "09:41" below the dial. Do not use scripts, external assets or embedded images. CPU protocol: Cycles CPU, 2 threads. Full private source inputs remain archived locally; source links and hashes identify them.
This is the evaluation adaptation and a rerun guide, not a verbatim transcript of the full original request. Exact original inputs remain archived locally.
Original evaluation input SHA-256: e237e133ad6988ea13a95c43e5bdab8b286d2c4670a8e42e9e24d8875441dedc
Requests / complete responses: 7 / 7
Model session finished normally; first SVG and revisions are preserved, with semantic checks reported separately.
13 · SVG · Requested effort: Claude Code requested adaptive
Shared zero baseline, proportional bar heights, values, total and best quarter pass independent checks.

Actual SVG render from this case
The model SVG is rasterized and encoded as WebP for display; source SVG is not rewritten.
Original SHA-256: bd10e6fca8fddb1dfd9ebfe7b34e5b8c7f4cb4f5b67d961782d9f1d2593c4194
Public SHA-256: 9d55e7e5d53aa465fa95ca4f886971d243768584c55342a5cf2e4de1b7734012
These are qualitative AI observations, not independent human scores. Automated checks are listed separately.
Task source and rerun specification
Independent offline generation preserving task requirements; Claude Code tool loop and Cycles CPU differ from the source benchmark.
The linked creator or official example is a source. The output above is our independently generated run.
Data chart and proportional bars. Source inputs remain archived locally and are identified by hashes.
Return only a complete, self-contained static SVG with viewBox="0 0 800 600". Design a polished dashboard card titled "Quarterly Revenue". Draw a vertical bar chart for Q1=12, Q2=24, Q3=18, Q4=30, in that left-to-right order. All four bars must share a zero baseline, and bar heights must be exactly proportional to the values. Show a y-axis labeled 0, 10, 20, 30 and show the value above each bar. Add summary labels "Total: 84" and "Best: Q4". Use a light background and readable typography. Do not use scripts, external assets or embedded images. CPU protocol: Cycles CPU, 2 threads. Full private source inputs remain archived locally; source links and hashes identify them.
This is the evaluation adaptation and a rerun guide, not a verbatim transcript of the full original request. Exact original inputs remain archived locally.
Original evaluation input SHA-256: 882c407247829aa8ef726fef5bcfeb8f9a4df8bbe701974d065ce807e5b64a14
Requests / complete responses: 21 / 21
Model session finished normally; first SVG and revisions are preserved, with semantic checks reported separately.
14 · SVG · Requested effort: Claude Code requested adaptive
All six fixed Chinese strings appear once, with a readable hierarchy and recognizable bicycle icon.

Fixed six-string Chinese poster test
The model SVG is rasterized and encoded as WebP for display; source SVG is not rewritten.
Original SHA-256: 1d705628971d3ec35bef94355da1bf1ca0e3902f88c73d734ed2b300162a177d
Public SHA-256: 97aff205866557fe037259f2cfbe4bc60f81409b2ff5dfcb7925bf490940cddb
These are qualitative AI observations, not independent human scores. Automated checks are listed separately.
Task source and rerun specification
Independent offline generation preserving task requirements; Claude Code tool loop and Cycles CPU differ from the source benchmark.
The linked creator or official example is a source. The output above is our independently generated run.
Chinese poster and layout constraints. Source inputs remain archived locally and are identified by hashes.
只输出一个完整、自包含的静态 SVG,viewBox="0 0 800 600"。设计一张清晰、精致的中文活动海报。必须且仅包含以下六段文字,每段出现一次,不能增删字或改写:主标题“周末骑行计划”;副标题“慢一点,看见城市”;时间“10月24日 09:00”;地点“滨江公园东门”;说明“自备头盔 · 全程12公里”;按钮文字“立即报名”。画一个简洁的自行车图标。主标题最大,时间地点次之,说明最小。背景浅色,留足边距,所有文字保持在画布内,不重叠,不使用脚本、外部素材或嵌入图片。 CPU protocol: Cycles CPU, 2 threads. Full private source inputs remain archived locally; source links and hashes identify them.
This is the evaluation adaptation and a rerun guide, not a verbatim transcript of the full original request. Exact original inputs remain archived locally.
Original evaluation input SHA-256: 6a1d4abaaf4e555ed8610782705bf065c14630467dd4dc8436079f6c56d44655
Requests / complete responses: 13 / 13
Model session finished normally; first SVG and revisions are preserved, with semantic checks reported separately.
15 · SVG animation · Requested effort: Claude Code requested adaptive
The four-second SVG renders and moves, but legs do not remain attached to the pedals.
Download MP4 · Playback starts only when you press play.
Four-second capture of the original SVG at 10fps; feet detach from pedals and source is not repaired.

Actual SVG render from this case
The model SVG is rasterized and encoded as WebP for display; source SVG is not rewritten.
Original SHA-256: 7d58cea1377d42cf0b56c7ab35be478cb08550a7b559d9e684ae8e41163e45e5
Public SHA-256: 2408206842ca1c54c049303848f48e092d1eec59f983a13c552d127039005f95
These are qualitative AI observations, not independent human scores. Automated checks are listed separately.
Task source and rerun specification
Independent offline generation preserving task requirements; Claude Code tool loop and Cycles CPU differ from the source benchmark.
The linked creator or official example is a source. The output above is our independently generated run.
Animated cycling pelican. Source inputs remain archived locally and are identified by hashes.
Return only a complete self-contained animated SVG with viewBox="0 0 800 600". Draw a recognizable pelican riding a bicycle: long bill and throat pouch, two wheels, connected frame, saddle and handlebar. Animate rotating wheel spokes and a rotating crank with pedaling legs that remain attached to the rider and pedals. Use SVG SMIL animateTransform/animate/animateMotion only (no CSS animation), with a seamless 4-second repeating loop. Keep the whole bicycle and bird on the canvas. Use a light background. No scripts, foreignObject, external assets, external fonts or embedded images. Use only local SVG shapes and text. The animation must visibly move, not just contain animation tags. CPU protocol: Cycles CPU, 2 threads. Full private source inputs remain archived locally; source links and hashes identify them.
This is the evaluation adaptation and a rerun guide, not a verbatim transcript of the full original request. Exact original inputs remain archived locally.
Original evaluation input SHA-256: 54ff24c0b7a505c1b309f411ef8d8fd1741cb633219cbe6ea179fb7e6ae85dd9
Requests / complete responses: 8 / 8
Model session finished normally; first SVG and revisions are preserved, with semantic checks reported separately.
Download the evaluation dataset and source index