TASK 15 / HTML · V1 · Productivity tool

Book-Tracking App

读书追踪应用

4 runs · 15.A.01 · HTML

Compare works →
TASK 15 · 15.A.01 · HTMLBook-Tracking App
Open independent copy ↗
Book-Tracking App

Preview note: Embedded storage and downloads can be restricted by browser sandboxing; use the standalone view for full behavior.

ORIGINAL PROMPT

Original prompt

PROMPT/Phase1_TEST_PROMPTS_BILINGUAL.md · The original prompt is kept verbatim

Source ↗
ORIGINAL PROMPT

Original prompt

Build an app for keeping track of books.
TRANSLATION

Chinese translation

构建一个用于追踪书籍的应用。

Expected outcome

Since the prompt does not constrain platform or scope, a reasonable expected outcome is a working personal-library and reading-progress application. At minimum, users should be able to add, view, edit, and delete books and record title, author, and reading status. A fuller implementation may include page progress, start and finish dates, ratings, tags, favorites, search, filters, sorting, statistics, and local persistence. Status transitions must keep data coherent—for example, finishing should set the finish date and final progress, while reopening should resolve conflicting fields. If import/export exists, data should round-trip and avoid silent duplication. The UI should clearly handle an empty library, zero search results, invalid input, and corrupt storage.

Additional verification guidance

  • Complete add, view, edit, delete, undo where available, and reload-recovery flows, confirming field values, list order, and persisted state.
  • Exercise transitions among want-to-read, reading, paused, finished, and abandoned states, checking current page, total pages, start/finish dates, and progress-log coherence.
  • Combine search, filtering, and sorting with case, whitespace, tags, and no-result queries, confirming stable counts and no data loss.
  • Validate rating ranges, dates, page bounds, duplicates, corrupt storage, and import/export round trips; verify keyboard and narrow-screen usability.
THREE INTERPRETATIONS

Same task, three model interpretations

All three results use the same width and the same preview ratio; no model is marked as best.

COMPARISON METRICS

Comparison metrics

Cross-model comparison must note differences in harness, budget and isolation.

MetricDeepSeek-V4.1-Flashrun-deepseek-v4-1-flash-exp-0910-task-15-r1Muse Spark 1.3run-muse-spark-1-3-task-15-r1GPT-5.6 Solrun-gpt-5-6-sol-task-15-r1K3run-k3-task-15-r1
Runtime14:261:356:529:06
API calls7519828
Tool calls7421627
Failures / retries4003
Token throughput7.1M411.9K273.7K1.1M
AI review99/10080/10079/10090/100
RUN RECORDS

Run records

Two runs of the same task version can be compared side by side.

15.A.01 · run-deepseek-v4-1-flash-exp-0910-task-15-r1DeepSeek-V4.1-Flash● Archived
Runtime
14:26
Token throughput
7.1M
API calls
75
View archive record →
15.B.01 · run-muse-spark-1-3-task-15-r1Muse Spark 1.3● Archived
Runtime
1:35
Token throughput
411.9K
API calls
19
View archive record →
15.C.01 · run-gpt-5-6-sol-task-15-r1GPT-5.6 Sol● Archived
Runtime
6:52
Token throughput
273.7K
API calls
8
View archive record →
RELATED TASKS

Related tasks

View all tasks →