ChatGPT and Claude in Film Reviews: Narrative and Technical Analysis
A practical same-input test plan for comparing ChatGPT and Claude on film reviews, grounded in vendor-documented file limits and accuracy cautions.
Sources checked 2 Oct 2026
OpenAI documents file analysis, including PDFs, and Anthropic documents text and visual analysis for some uploaded PDFs, but neither vendor page establishes which assistant gives the stronger film review. Both vendors warn that outputs can be wrong or misleading, so you need identical inputs and an evidence check rather than a confident tone. Chat Picker has not tested, scored, or ranked the assistants for this use.
Know these limits before you start
Build a compact packet from files both services accept. As read on October 1, 2026, OpenAI's file-uploads FAQ gives a 512 MB hard limit per file, a 2M-token cap for each text or document file, and a 20 MB limit per image. Anthropic's upload page, also read on October 1, 2026, lists 500 MB per chat file and 30 MB per project file. It says Claude analyzes text and visual elements in PDFs of 100 pages or fewer, but only text in PDFs from 101 to 1,000 pages. These are handling rules, not measures of review quality.
Check the data setting before using unpublished material. OpenAI's pricing page, read on October 1, 2026, says an opt-out is available on Free, Go, Plus, and Pro. Anthropic's pricing page, read on the same date, says the opt-out applies to Free, Pro, and Max, while Team is not trained on by default. The Google AI Plans page read that day does not state a Gemini app training setting, so Chat Picker marks that cell Not verified.
Treat both outputs as drafts. OpenAI's accuracy notice says ChatGPT can be incorrect or sound confident when wrong and urges source checks. Anthropic's accuracy notice warns that Claude can produce convincing but ungrounded quotations and says not to rely on it as the only source of truth. For consumer-facing output, Anthropic's Usage Policy requires AI disclosure at the beginning of a session. OpenAI's Usage Policy prohibits inferring an individual's emotions in workplace and educational settings except for medical or safety reasons; keep this test on fictional characters.
The test plan
Run each pass in a fresh chat, give both assistants the same packet and prompt, and save the full response. Record the plan and account setting with every answer.
-
Sentiment and narrative reasoning. Prepare a review excerpt and a plot synopsis covering the same film. Give this instruction: “Classify each claim as sentiment, plot inference, or technical observation, explain the evidence, and place any claim not supported by the synopsis under Needs checking.” Look for unjustified labels and narrative claims mixed with interpretation. Record the label, supporting wording, and each unsupported claim.
-
Plot structure. Prepare a synopsis with its scenes in story order. Give this instruction: “Map the plot structure. Mark every event as directly stated or inferred, attach a short supporting quotation to each stated event, and flag any ordering claim that conflicts with the synopsis.” Look for invented events and sequence errors. Record support status and every conflict.
-
Character arcs. Prepare the synopsis and notes recording each character's stated starting goal, ending goal, and turning event. Give this instruction: “Make a character arc table, and mark each internal-state claim as stated or inferred; distinguish a change in belief from a change in circumstance.” Look for feelings or motives assigned without textual support. Record the evidence status for each arc claim.
-
Theme and subtext. Prepare a review passage and separate notes for its explicit theme. Give this instruction: “Separate explicit theme claims from subtext claims. Quote the relevant wording and label each subtext claim supported, contradicted, or unresolved.” Look for subtext presented as fact. Record the quotation, label, and any contradiction.
-
Pacing and time. Prepare scene-timing notes and a review excerpt containing temporal wording. Give this instruction: “Separate observed time changes from inferred effects on tempo, quote the review's temporal vocabulary, and flag any timestamp or sequence claim absent from the notes.” Look for invented timing details and generic pacing claims. Record observed, inferred, and unsupported items.
-
Comparative context. Prepare multiple review excerpts with their corresponding source notes. Give this instruction: “For every shared claim, label it direct support, interpretation, conflict, or absent; quote the source wording and introduce no outside facts.” Look for merged sources, misattribution, and unsupported cross-references. Record the cited source, claim status, and any conflict.
How to read your results
Read the log by evidence failures, not polish. Reject publishable copy containing an invented quotation, plot event, source attribution, or timestamp. Next review unsupported motives, themes, and causal links.
If both assistants avoid those errors, choose by workflow fit: documented file limits and data settings that match your packet and process. Repeat the decisive pass with another source before treating one response as the better fit. The log describes your run, not a general ranking.
Where the plans differ
In the Claude vs ChatGPT vs Gemini matrix, checked on October 2, 2026, start with the rows for Free plan, Main paid plan, Context window in the app, How usage limits are described, Training on your chats, and Ads. Then check the linked file documentation for your packet.
As read on October 1, 2026, OpenAI's Plus help page lists Plus at $20 a month; Anthropic's pricing page lists Pro at $20 month to month or $17 a month with annual billing and $200 up front; and Google's US plans page lists AI Pro at $19.99 a month. None of those prices documents review quality.
On the same pages, OpenAI lists app context windows by model and plan, while Anthropic says up to 1M on every plan, varying by model; Google gives no app token figure. OpenAI and Anthropic describe usage without publishing message counts on those pricing pages. OpenAI also says Go may include ads, while the Claude and Google pricing pages do not state app ad terms. Confirm the live vendor page before subscribing because plans and limits can change.