AI CODING ASSISTANTS: 30-DAY SPEND AND OUTPUT CHECK Heresthebest.com | Verified plan pages 6 November 2026 | Reader-run worksheet MY EDITOR / REPOSITORY: ____________________ MY REAL JOB TO TEST: fix one reproducible bug, create 3 tests and prepare a PR. START DATE ____ END DATE ____ BILLING COUNTRY/CURRENCY ____ CURRENT APPROVED WORKFLOW (e.g. tests/build/review) __________________ PLAN 1: GitHub Copilot Official US prices at review: Free $0; Pro $10; Pro+ $39; Max $100/month. AI Credit monthly allowances: Pro 1500, Pro+ 7000, Max 20000 (flex usage may vary). Paid completions are not the same as AI Credits. https://github.com/features/copilot/plans Chosen plan ____ base quote ____ spend ceiling ____ overage ____ TOTAL ____ PLAN 2: Cursor for NEW subscribers Official Nov 2026 page: Free $0; Plus $8; Premium $30; Super $100; Ultra $200 per month. Cursor is now bundled with the X Subscription. Plus supports Grok models; third-party models are separately metered at API list price on Premium and higher. Older Pro/Pro+/Ultra plans are LEGACY, not the new buyer price ladder. https://cursor.com/pricing https://cursor.com/docs/models-and-pricing Chosen tier ____ tier fee ____ THIRD-PARTY prepaid credit used ____ TOTAL = tier fee + actual separately metered usage ____. PLAN 3: Devin Desktop (formerly Windsurf) Official examples: Free $0; Pro $20/month; Max $200/month. Teams: $80/mo team plan plus $40/mo per full developer seat. https://devin.ai/pricing Chosen plan ____ seats ____ agent extras ____ TOTAL ____. PLAN 4: Claude Code Claude Pro: US $20 billed monthly or $200 billed annually (about $17/mo). Claude Max starts at $100/mo, with usage limits. Subscription usage for Claude and Claude Code is shared; API billing is separate. https://claude.com/pricing https://code.claude.com/docs/en/costs Plan cost ____ actual API/extra spend ____ TOTAL ____. HYPOTHETICAL MATH ONLY (NOT ACTUAL USAGE) Copilot Pro $10 plus assumed $0 extra = $10. Cursor Premium $30 plus assumed $18 third-party model usage = $48. Difference $38, but the products, models and available quotas differ. SAME TASK EVALUATION (do this in YOUR authorized repo) Baseline minutes without assistant ____ With plan A: minutes ____ tests pass? __ PR reviewed? __ With plan B: minutes ____ tests pass? __ PR reviewed? __ Did the agent change files outside the request? __ Could you safely reproduce the result and roll it back? __ Data/privacy & repository permission requirements satisfied? __ BREAK-EVEN Time saved (minutes) ____ / 60 * dollar value per hour ____ = $____. Less total incremental subscription and usage spend = NET $____. Do NOT interpret hypothetical time savings as realized cash revenue. IMPORTANT STOP RULES - Verify plan page, country, checkout amount and model access before payment. - Do not equate 'unlimited autocomplete' with unlimited agent use. - Do not buy or migrate based solely on an outdated Cursor Pro price table. - Set API/prepaid credit alerts and a monthly cap before a big agent task. - Confirm third-party models, private repo access and data retention terms. - Cancel or revert if tests, review quality or security deteriorate.