FreshLensIAS · All GS Papers · OpenAI Bench

Four papers, one desk

Select the GS paper; its Holistic Final Evaluator loads as the system prompt (GS1, GS2, GS4 on v1.0; GS3 on Rubric v2, WCI system), and its ladders, count ceilings, and gates are enforced in code. Everything else is common: two files in (question paper optional), parallel per-answer evaluation, evidence-locked reports, per-mentee markdown download.

Setup

Checking sign-in…

Each evaluated answer uses 1 credit. Credits are assigned by FreshLensIAS; contact your mentor to top up.

General Studies paper

Evaluation model

Or exact model string (overrides dropdown; priced at dropdown rates)

Parallel calls: Higher is faster; drop to 1 or 2 if a new account keeps hitting rate limits. A 20-question copy at 3 finishes in roughly 8 to 15 minutes.

System prompt · loads from the selected paper (editable per session)

The selected paper's rubric verbatim, plus the API preamble and JSON schema. Prose edits survive paper-switching within the session; changing tables or thresholds also requires editing PAPERS in this file's script.

Submissions

Question paper OPTIONAL

One PDF. Leave empty if questions are written above each answer in the copy.

Answer copy

One PDF, the full booklet scanned in one go (up to 20 questions, under 22 MB). Selecting again replaces it.

Copy-level report

Total awarded / maximum

–

Q	Question	Marks	Category	Status

Question-wise detail (tap to expand; internal panel inside)

Four papers, one desk

Token consumption & cost · this whole copy

Sign in