AI Chat in Job Interview Prep: Mock Interviews and Resume Optimization
All three assistants offer free plans, while their documented paid prices and upload limits differ, so this afternoon test uses the same inputs and a personal decision log.
Sources checked 3 Oct 2026
For mock interviews and resume optimization, the OpenAI data-analysis page, Claude file-upload page, and Gemini file-upload page show that each assistant can work with files, but those vendor pages do not establish which gives the best result. Chat Picker has not tested, measured, scored, benchmarked, or ranked the assistants for this use. This page offers an afternoon test and a decision log instead.
Know these limits before you start
Start with files. Here are the upload limits as read on October 1, 2026. OpenAI’s File Uploads FAQ gives every file a 512MB hard limit, caps each text or document file at 2M tokens, and allows Free accounts three uploads per day. Claude’s upload page lists 500MB per chat file and 30MB per project file. Gemini’s file-upload page allows up to 10 supported files per prompt, subject to availability; non-video files can be 100MB, and a work or school administrator can block Drive access.
Check data controls before uploading a current resume or private work document. The OpenAI pricing page says an opt-out is available on Free, Go, Plus, and Pro. The Claude pricing page says the same for Free, Pro, and Max, and that Team is not trained on by default. The Gemini plans page does not state a training setting for chats, so Chat Picker’s comparison matrix marks it “Not verified.”
Build verification into the session. OpenAI’s accuracy note says ChatGPT can be incorrect or misleading and may sound confident. Anthropic’s note says Claude can make convincing but ungrounded statements and should not be the only source of truth. Google’s Gemini guidelines describe model outputs as probabilistic and acknowledge limited training data and overgeneralization. Check claims against the original files and sources. OpenAI’s usage policies prohibit automated high-stakes decisions in sensitive areas without human review; Anthropic’s Usage Policy makes users and organizations responsible for accuracy and calls for relevant human expertise in elevated-risk uses.
The test plan
Use one spreadsheet row per run: assistant, plan, task, input, completion, instruction status, problem, evidence, next action, and displayed charge.
-
Behavioral mock interview. Prepare one current resume, one job description for a product operations manager, and three brief STAR notes. Say: “Act as an interviewer for the product operations manager role. Ask one behavioral question at a time. Begin, ‘Tell me about a time you handled conflicting priorities.’ After each answer, ask one relevant follow-up, then wait.” Log question relevance, follow-up quality, and ignored instructions.
-
Resume and ATS-style keywords. Use the same files. Say: “List exact job-description phrases, mark whether the resume supports each one, and suggest truthful rewrites for gaps. Add no skill, employer, date, credential, or result. Quote every edit and its location.” Log invented content, ambiguous edits, and formatting lost after upload.
-
Feedback depth. Attach one answer and your notes. Say: “Separate demonstrated facts from implied claims. Give one wording revision and one rehearsal drill, quote the evidence, and write ‘not verifiable’ when the files do not support a suggestion.” Record specificity, priority, and usability.
-
Persona retention. In one continuous chat, say: “Act as a hiring manager for the product operations manager role. Ask five behavioral questions one at a time and ask one relevant follow-up after each. Do not break role or encourage me. After the fifth answer, summarize concerns without making a hiring decision.” Log tone drift, contradictions, repeated questions, and role breaks.
-
Cost and access. Open the account and plan pages. Say: “List the file types available in this chat and any interface limit. Do not guess a number, then confirm that you can read the attached resume and job description.” Repeat the mock interview. Log plan, billing cadence, upload result, cap messages, tool access, displayed charge, and time used. Do not estimate an unshown limit.
-
Verdict and model check. Attach the documents and notes. Say: “Separate required fixes from optional suggestions. For each required fix, identify the exact passage or missing evidence and propose a truthful revision. State what you cannot verify. Recommend which available mode to repeat, cite supporting log entries, and do not claim a universal best.” Log unresolved advice and unsupported conclusions.
How to read your results
Read row by row, not by overall polish. Apply hard gates first: the real files work, the core instruction is followed, the session fits your plan, and no data or policy concern is unresolved. Then look for traceable edits, feedback that separates evidence from inference, and a stable persona.
Give more weight to a problem repeated across tasks than to one polished answer. Keep access failures separate from writing quality. If a mode recommendation lacks a traceable log entry, mark it unsupported. Choose the workflow that meets your needs, or revise the prompt and rerun. Do not turn one session into a general ranking.
Where the plans differ
Before paying, open Chat Picker’s Claude vs ChatGPT vs Gemini matrix. Focus on Free plan, Main paid plan, Context window in the app, How usage limits are described, Training on your chats, and Ads. Recheck vendor pages because plans and regional prices change.
- Free plan: Each vendor lists free access, as read on October 1, 2026. ChatGPT limits uploads, images, voice, and deep research; Claude lists web search, file creation, code execution, and memory; Gemini requires a Google Account and lists 15GB storage. See OpenAI pricing, Claude pricing, and Google AI plans.
- Main paid plan: ChatGPT Plus is $20/month. Claude Pro is $20 monthly or $17/month with annual billing and $200 up front. Google AI Pro is $19.99/month. These US-dollar figures were read on October 1, 2026 from the linked vendor pages.
- Usage and context: OpenAI and Claude publish no message counts on their pricing pages. Google says compute limits refresh every five hours, up to a weekly limit, with paid tiers at 2x, 4x, and up to 20x. The context row lists plan-specific OpenAI values and up to 1M for Claude on every plan; Gemini gives no app token figure, so its cell remains blank. These figures were read on October 1, 2026.
- Training and ads: Use the training controls above. ChatGPT’s Go plan “may include ads”; the Claude and Gemini pages read do not state an ads policy. The Go wording was read on October 1, 2026 from OpenAI pricing. For heavier use or shared access, inspect Heavy-use plans and Team plan; do not extrapolate from personal prices.