Every tool here has a free tier, and they hand out very different amounts of real agent work. GitHub Copilot's free plan is the only one that includes CLI and agent mode with no credit card, though chat and agent usage are capped. Kiro gives 50 credits a month with open-weight models and Claude Sonnet 4.5, which covers two or three real feature builds at 15 to 25 credits a spec run. Cursor's Hobby tier gives limited agent requests and Composer access with no card required. Windsurf gives a light daily agent quota plus unlimited tab completions and limited model availability. Claude Code's free Claude plan is chat rather than a coding agent, so it does not belong on a free-agent shortlist even though the brand is free. The paid step-up is where the long-term cost shows: Copilot Pro at $10 is the cheapest real seat, Claude Code Pro at $17 on annual is next, and Cursor, Windsurf and Kiro all enter at $20. Watch the meters that bill separately: Copilot's credit pool adds flex allotments of $5 to $100, Cursor moves to usage-based overage above its request caps, Kiro sells add-on credits at $0.04, and Claude Code bills API tokens at $2 input and $10 output per million. Decision: stay free on Copilot or Kiro for real agent work, graduate to Copilot Pro for the cheapest paid seat, and pick Claude Code's API when volume beats a flat fee. Prices re-checked on the date shown.
4.4/5: Free $0 includes CLI and agent mode with limited chat and agent usage; Pro $10 a month is the cheapest paid lift in this comparison.
Free tier: Yes
Paid from: $10/mo (Pro) (Usage is metered on a two-track credit system: each plan carries base credits (Pro $10/mo, Pro+ $39/mo, Max $100/mo) plus a variable flex allotment (Pro $5, Pro+ $31, Max $100), so the Pro+ total reaches $70 and the Max total $200 in monthly AI credits. One AI credit equals $0.01. Pooled organization credits are included on team tiers.)
GitHub Copilot is the IDE-native default: free for personal use, Pro at $10 a month, Pro+ at $39 and Max at $100, with a separate credit meter for agent usage on top of the plan price.
3.9/5: Free 50 credits a month with Claude Sonnet 4.5 and open-weight models, 14-day 500-credit bonus for new signups, Pro $20 lifts it to 1,000 credits.
Free tier: Yes · New signups get a 14-day 500-credit bonus; first paid upgrade via social login or AWS Builder ID credits $20
Paid from: $20/user/mo (Pro) (Kiro absorbed Amazon Q Developer's IDE role (Q Developer closed to new signups on 2026-05-15, end of support 2027-04-30). Credits are a unit of agent work, not a single interaction: a full spec run consumes 15 to 25 credits, add-on credits cost $0.04 each, and paid tiers can pool overages for teams.)
Kiro is AWS's spec-driven IDE: free with 50 credits a month, then Pro at $20, Pro+ at $40, Pro Max at $100 and Power at $200 per user, with overage at $0.04 per credit.
4.5/5: Hobby $0 with no credit card, limited agent requests and Composer access; Individual $20 a month lifts the caps and adds frontier models and cloud agents.
Free tier: Yes
Paid from: $20/mo (Individual) (Subscription tiers carry limits on agent requests, frontier-model access and cloud agents; overage is billed on a usage basis. Teams adds centralized billing, SSO, team marketplaces and a team-wide privacy mode; Enterprise adds pooled usage, SCIM and audit logs.)
Cursor is the agentic editor: a free Hobby tier, Individual from $20 a month across Pro, Pro+ and Ultra, Teams at $40 per user, and usage-based billing above the limits.
4.0/5: Free light daily agent quota with unlimited tab completions and limited model availability; Pro $20 a month lifts the quota and unlocks SWE-1.5 and Claude/GPT.
Free tier: Yes
Paid from: $20/mo (Pro) (Pricing is quota-based, not credit-based: each tier carries daily and weekly agent quotas and access to model families including SWE-1.5, Claude and GPT. Cloud agents and CLI access extend to Pro through a dated promotion, and Extra Usage is bought at API pricing above the quota.)
Windsurf, now part of Cognition, is the quota-priced editor: free with a light agent quota, Pro at $20, Max at $200, and Teams at $80 plus $40 per full developer seat.
4.5/5: the free Claude plan is chat, not a coding agent; Pro $17/mo annual ($20 monthly) or the API at $2/$10 per MTok is the entry to Claude Code itself.
Free tier: Yes
Paid from: $17/mo annual (Pro) (Claude Code is reachable three ways: bundled into a Pro or Max subscription, as a team seat, or API-only at per-token rates (Sonnet 5.5: $2 input / $10 output per million tokens; Opus 5.5: $4 / $20), which is cheaper than any flat seat once usage is high.)
Claude Code is the model-native terminal agent: it runs inside Claude Pro at $20 a month, Claude Max from $100, or straight off the API at per-token rates with no seat fee.
Bottom line: GitHub Copilot has the only free tier that genuinely runs an agent, Kiro's 50 credits are the clearest free allowance for spec-driven work, Cursor and Windsurf give free editors with limited or daily-quota agent access, and Claude Code has no free agent tier at all. Free tiers are for proving the workflow — the cap lifts at $10 on Copilot, $17 on Claude Code, and $20 on Cursor, Windsurf and Kiro.
GitHub Copilot, whose free plan includes CLI and agent mode. Kiro's 50 credits run spec-driven agent work, Cursor's Hobby tier gives limited agent requests, and Windsurf's free tier gives a light daily quota. Claude Code's free plan is chat only.
Not for coding. The free Claude plan is chat on web, desktop and mobile; Claude Code itself starts at Pro, $17 a month billed annually or $20 monthly, or through the API at $2 input and $10 output per million tokens for Sonnet 5.5.
About two to three feature builds, since a full spec run consumes 15 to 25 credits. New signups also get a 14-day 500-credit bonus, and the first upgrade to a paid plan credits $20 via social login or AWS Builder ID.
GitHub Copilot Pro at $10 a month. Claude Code Pro is $17 on annual billing, and Cursor, Windsurf and Kiro each enter at $20.
Windsurf includes unlimited tab completion on its free tier; Copilot's free plan includes completions with usage caps on chat and agent features.
Yes: Copilot's credit pool carries flex allotments of $5 to $100, Cursor bills usage-based above its agent-request caps, Kiro sells add-on credits at $0.04, and Claude Code's API bills per token.