On this page
AI coding tools change their prices more often than any other software I pay for. Cursor swapped 500 requests for a dollar pool in June 2025 and apologised 18 days later. GitHub moved every Copilot plan onto token-metered credits on 1 June 2026. Windsurf went the other way, dropped credits for daily quotas, then stopped being called Windsurf. Most comparison pages are a snapshot from whenever someone last wrote them.
This page is a tracker instead. It lists what each tool costs today, what the limit on each plan actually is, and what happens when you hit it. Below that sits a dated log of every pricing change I could source, and the weekly page reads that catch the next one.
- $20 a month is the median entry price of an AI coding tool, across the 14 of 16 tools that publish a self-serve paid plan (entry prices mix one-person, per-seat and whole-team plans; each is labelled in the table); counting one-person plans only, the median is $20 across 11 tools. Claude Code, Cursor, Windsurf/Devin Desktop, Kiro and Replit all open at exactly $20 a month for one person.
- OpenAI Codex via ChatGPT Go at $8 a month is the cheapest paid plan in the tracker, ahead of GitHub Copilot Pro and JetBrains AI Pro at $10. Cursor's India-only Start plan (INR 649 a month) and the free tiers sit outside this comparison.
- 9 of 15 tools with published pricing meter usage in credits, tokens or a dollar pool; 4 use rolling time-window quotas (Claude Code, Windsurf/Devin Desktop, OpenAI Codex and Google Jules), 1 still counts requests (Amazon Q Developer) and 1 bills every token to your own key (Cline).
- 5 tools moved their limits onto a usage meter since June 2025 (GitHub Copilot, Kiro, JetBrains, Cursor, Replit); 1 did so in the last 12 months (GitHub Copilot), while 1 moved the other way (Windsurf/Devin Desktop, from credits to quotas).
- 6 tools now sell an individual plan at $199.99 a month or more, and 2 pricing changes landed in the last 90 days. The change log holds 26 dated changes from 11 tools since April 2025.
Updated . Data CC BY 4.0; download the plans CSV and the change log CSV.
Cite this page: "AI coding tools pricing tracker", Gravity, updated , https://gravity.fast/data/ai-coding-tools-pricing/. Data CC BY 4.0: copy the rows into your own work and link back here. Downloads: plans CSV and change log CSV.
Current prices: 16 AI coding tools
One row per tool, US list prices before tax. "What the limit is" says which meter runs out first, because two $20 plans can buy very different amounts of work. The last-change date links to the row in the change log.
| Tool | Free tier | Paid plans (US dollars) | What the limit is | Past the limit | Last change | Checked |
|---|---|---|---|---|---|---|
| Claude Code Anthropic | Claude Free does not include Claude Code. | Individual: Pro $20 a month ($17 billed annually); Max 5x $100 a month, 5x Pro usage per 5-hour session; Max 20x $200 a month, 20x Pro usage per 5-hour session Team: Team Standard $25 a seat a month ($20 billed annually); Team Premium $125 a seat a month ($100 billed annually), 5x a Standard seat; Enterprise $20 a seat a month billed annually | Rolling time-window quota. Rolling 5-hour session limit plus weekly limits, one pool shared with Claude chat. Max gives 5x or 20x Pro usage per 5-hour session; Team Premium seats give 5x a Standard seat. | Optional usage credits at standard API rates on paid plans; or pay per token through the API (Opus 5.5 $4 in / $20 out, Sonnet 5.5 $2 / $10, Haiku 4.5 $1 / $5 per million tokens). | by 29 January 2026 | 4 October 2026 |
| Cursor Anysphere | Hobby: free, limited Agent requests. | Individual: Pro $20 a month; Pro Plus $60 a month; Ultra $200 a month; Start (India only) INR 649 a month, tax inclusive Team: Teams Standard $40 a seat a month; Teams Premium $120 a seat a month, 5x the Standard Agent limits | Credits, tokens or a dollar pool. Each plan includes a monthly amount of model usage in two pools: Cursor's own models (Grok, Composer) and third-party models charged at their API prices. | On-demand usage at the same API rates, billed in arrears. Teams and Enterprise add a Cursor Token Rate of $0.25 per million tokens on third-party models. | 28 July 2026 | 4 October 2026 |
| GitHub Copilot GitHub (Microsoft) | Copilot Free: 2,000 completions a month plus a small AI Credits allowance, Auto model selection only. Free for verified students. | Individual: Pro $10 a month, 1,500 AI credits (1,000 base + 500 flex); Pro+ $39 a month, 7,000 AI credits (3,900 base + 3,100 flex); Max $100 a month, 20,000 AI credits (10,000 base + 10,000 flex) Team: Business $19 a seat a month, 1,900 AI credits per user; Enterprise $39 a seat a month, 3,900 AI credits per user | Credits, tokens or a dollar pool. Monthly GitHub AI Credits (1 credit = $0.01), drawn down by tokens at listed API rates. Individual plans get base credits plus a variable flex allotment. | Set a dollar budget and keep going at $0.01 a credit; on Business and Enterprise, admins decide whether paid overage is allowed. | 1 June 2026 | 4 October 2026 |
| Devin Desktop (formerly Windsurf) and Devin Cognition | Free: a light agent quota, unlimited Tab completions and inline edits. | Individual: Pro $20 a month; Max $200 a month Team: Teams $80 a month for the team plus $40 a month per full seat, Up to 200 users; each full seat has its own quota | Rolling time-window quota. A usage allowance that refreshes daily and weekly; cost per message varies with model and task size. | Buy extra usage, consumed at API pricing. | 2 June 2026 | 4 October 2026 |
| OpenAI Codex OpenAI | ChatGPT Free includes Codex for quick tasks. | Individual: Go $8 a month; Plus $20 a month; Pro $100, $200 or $500 a month, No 5-hour limit; Ultrafast on the $500 tier Team: Business $25 a seat a month ($20 billed annually), 2+ users; larger cloud VMs | Rolling time-window quota. Included in ChatGPT plans: a 5-hour usage window plus weekly limits (Plus: roughly 15 to 160 local GPT-6.1 Sol messages per 5 hours). Pro plans currently have no 5-hour limit. | ChatGPT credits on Plus, Pro and Business; or an API key billed per token. | none logged | 4 October 2026 |
| Google Jules, Antigravity and Gemini Code Assist | Gemini Code Assist for individuals and Antigravity for individuals are free with limits; Jules has a free tier. | Individual: Google AI Pro $19.99 a month; Google AI Ultra $99.99 (5x Pro) or $199.99 (20x Pro) a month Team: Code Assist Standard $22.80 a seat a month ($19 billed annually), 1,500 model requests a day; Code Assist Enterprise $54 a seat a month ($45 billed annually), 2,000 model requests a day | Rolling time-window quota. Jules: daily task limits on a rolling 24 hours (15 free, 100 with Google AI Pro, 300 with Ultra). Antigravity: weekly rate limits. Gemini Code Assist: daily model-request limits. | Antigravity adds a flexible AI credit pool on Pro and Ultra; Jules has no paid overage, new tasks wait for the window to reset. | by 3 June 2026 | 4 October 2026 |
| Amazon Q Developer Amazon Web Services | Free tier: 50 agentic requests a month. | Team: Pro $19 a seat a month | Request count. Agentic requests a month: 50 on Free, higher limits on Pro. Java upgrades: 4,000 lines of code a month per Pro user, pooled. | None for agentic requests; extra Java-upgrade code at $0.003 a line. | none logged | 4 October 2026 |
| Kiro Amazon Web Services | Free: 50 credits a month. | Individual: Pro $20 a seat a month, 1,000 credits; Pro+ $40 a seat a month, 2,000 credits; Pro Max $100 a seat a month, 5,000 credits; Power $200 a seat a month, 10,000 credits | Credits, tokens or a dollar pool. Monthly credits, charged fractionally by prompt complexity and model (a task on Sonnet 4.6 costs 1.3x the Auto agent). Unused credits do not roll over. | Add-on credits at $0.04 each (packs from $5), valid 12 months. | 11 June 2026 | 4 October 2026 |
| JetBrains AI (Junie) JetBrains | AI Free: 3 AI Credits every 30 days, plus your own API keys. | Individual: AI Pro $10 a month ($8.33 billed annually), 10 AI Credits ($100 a year billed annually); AI Ultimate $30 a month ($25 billed annually), 35 AI Credits ($300 a year billed annually) Team: AI Pro (organizations) $20 a seat a month, 20 AI Credits; AI Ultimate (organizations) $60 a seat a month, 70 AI Credits; AI Enterprise $60 a seat a month, At least the Ultimate quota ($720 a year) | Credits, tokens or a dollar pool. AI Credits every 30 days, 1 credit = $1 (AI Pro 10, AI Ultimate 35 for individuals). | Top up AI Credits in any amount; top-ups valid 12 months. | by 18 September 2025 | 4 October 2026 |
| Replit Agent Replit | Free Mode usage on paid plans; a free account exists with limited usage. | Individual: Core $20 a month ($18 billed annually), $20 toward the most capable models; Pro $100 a month ($90 billed annually), $100 toward the most capable models, 10 parallel agents | Credits, tokens or a dollar pool. A dollar allowance toward the most capable models each month, plus Free Mode usage; agent work is priced by effort per request. | Pay as you go at effort-based rates, with spend controls. | 24 February 2026 | 4 October 2026 |
| Lovable Lovable | Free: limited daily credits, unlimited collaborators. | Individual: Pro $25 a month, 100 credits, unlimited users Team: Business $50 a month, 100 credits, SSO, team workspace | Credits, tokens or a dollar pool. 100 monthly credits on Pro and Business, one balance for building, chat, Cloud and AI; credits roll over for one month. | On-demand credit top-ups. | June 2026 | 4 October 2026 |
| Bolt.new StackBlitz | Free: 300,000 tokens a day, 1 million a month. | Individual: Pro $25 a month, From 10M tokens a month Team: Teams $30 a seat a month | Credits, tokens or a dollar pool. AI tokens per month (Pro starts at 10 million); unused tokens roll over one month. | Buy more tokens or move to a larger token tier. | 1 July 2025 | 4 October 2026 |
| v0 Vercel | Free: 7 messages a day. | Team: Plus $30 a seat a month, $30 of credits per user; Business $100 a seat a month, $30 of credits per user, training opt-out | Credits, tokens or a dollar pool. Dollar credits per user each month, drawn down by tokens at v0 model rates (v0 Pro $2 in / $10 out per million tokens), plus $2 of daily login credits. | Buy additional credits, shared across the team. | none logged | 4 October 2026 |
| Augment Code Augment Code | No free plan listed. | Team: Standard $20 a month for the team, $20 of usage, up to 50 seats; Business $100 a month for the team, $100 of usage, up to 50 seats | Credits, tokens or a dollar pool. A monthly dollar pool of usage across models, the Context Engine and compute, shared by up to 50 seats. | Top-ups, pay as you go. | none logged | 4 October 2026 |
| Cline Cline Bot | The extension is free and open source for individuals. | Team: Enterprise Custom | Pay per token (bring your own key). No plan limit. You pay the model provider for every token, through your own API key or Cline's provider at cost. | Not applicable: every token is billed. | none logged | 4 October 2026 |
| Tabnine Tricentis | Not published. | No self-serve price published | Not published. Not published. The pricing URL redirects to a Tricentis sales contact form. | Not published. | 31 July 2026 | 4 October 2026 |
16 tools. Prices are list prices before tax, on monthly billing unless marked. "A seat a month" is per user; "a month for the team" is one price for a whole team; the rest are one person's plan. A seat price and a usage price are different units: compare what each plan includes, not the headline alone.
Claude Code, Cursor and Copilot pricing in plain terms
These three get searched more than the rest of the table put together, so here is each one in a paragraph. Figures are as of 4 October 2026; the table above updates first if they move.
Claude Code pricing
Claude Code has no subscription of its own. It comes with a Claude plan: Pro at $20 a month ($17 billed annually), Max at $100 for 5x Pro usage or $200 for 20x, and Team seats from $25 a month. Every plan runs on a rolling 5-hour window plus weekly limits, shared with ordinary Claude chat, and on paid plans you can switch on extra usage at API rates once you hit them. Anthropic's own cost documentation puts the average enterprise spend at about $13 per developer per active day, or $150 to $250 a month, with 90% of users under $30 a day.
The other route is the API, where you pay per token (Sonnet 5.5 at $2 per million input tokens and $10 per million output, Opus 5.5 at $4 and $20). For the history of those model prices, see the LLM API price history.
Cursor pricing
Cursor Hobby is free with limited agent requests. Pro is $20 a month, Pro Plus $60 and Ultra $200. Each paid plan includes two monthly pools: one for Cursor's own models and one for third-party models charged at their API prices. Past both, on-demand usage continues at the same rates and is billed in arrears. Cursor's own guide says daily agent users typically spend $60 to $100 a month in total. Teams is $40 a seat, or $120 for a Premium seat with 5x the agent limits, and developers in India can buy a Start plan for INR 649 a month that covers only Cursor's models.
GitHub Copilot pricing
Copilot Free costs nothing and includes 2,000 completions a month. Pro is $10 a month, Pro+ $39 and Max $100; Business is $19 a seat and Enterprise $39. Since 1 June 2026 each plan includes GitHub AI Credits worth $0.01 each, drawn down by tokens at API rates: 1,500 a month on Pro, 7,000 on Pro+ and 20,000 on Max, counting the flex allotment GitHub adds on top of the base. Past that, you set a dollar budget and keep going at the same rate.
The cheapest AI coding assistant
On paid plans, ChatGPT Go at $8 a month is the cheapest way into a coding agent (OpenAI Codex for lightweight tasks), then GitHub Copilot Pro and JetBrains AI Pro at $10. Several tools are free to start: Copilot Free, Gemini Code Assist for individuals, Cursor Hobby, Antigravity for individuals and Cline, which is free software where you pay the model provider directly through your own API key. Cheapest per month is not cheapest per task, though. At the six tools that put a dollar value on the entry plan's included usage (Copilot, Kiro, JetBrains, Replit, v0 and Augment), that usage is worth at least the plan price, either as stated in dollars or valued at the vendor's own overage rate, as of 4 October 2026. If you are weighing coding tools against general agent platforms, the cheapest AI agent platforms ranking covers the other side.
The change log
One row per change, newest first, each with a source. "Limit unit" shows what the plan was metered in before and after, for example requests to credits. "Type" is my label: an increase or decrease means you pay more or less for the same thing; a restructure changes what the plan includes; a new plan adds a tier.
| Date | Tool | What changed | Old → new | Limit unit | Type | Source |
|---|---|---|---|---|---|---|
| Tabnine Tabnine | Tricentis acquired Tabnine. By 4 August 2026 tabnine.com/pricing redirected to a Tricentis sales form, so no self-serve price is published. | Public per-user price list → Contact sales | seat | Price withdrawn | ChannelLife, Wayback Machine (tabnine.com/pricing, 4 Aug 2026) | |
| Cursor Cursor Start (India) | Cursor launched an India-only plan priced in rupees and payable by UPI, covering only Cursor's own models. | Cheapest paid individual plan: Pro at $20 a month → Start at INR 649 a month, tax inclusive | credits | New plan | Cursor, Cursor Docs | |
| Kiro Kiro Pro Max | Kiro added a $100 Pro Max tier between Pro+ and Power. | Pro+ $40 (2,000 credits), then Power $200 (10,000 credits) → Pro Max $100 a month with 5,000 credits | credits | New plan | Kiro | |
| Google Jules, Antigravity and Gemini Code Assist Google AI Ultra | Google split AI Ultra into two cheaper tiers. Ultra includes Jules in Ultra and the highest Antigravity limits. | Google AI Ultra at $249.99 a month → $99.99 a month (5x Pro usage) or $199.99 a month (20x Pro usage) | quota | Decrease | Wayback Machine (gemini.google, 1 May 2026), Wayback Machine (gemini.google, 3 Jun 2026) | |
| Devin Desktop (formerly Windsurf) and Devin Devin Desktop (formerly Windsurf) | Windsurf became Devin Desktop and its plans merged into Devin's price list: the team plan now carries a base fee on top of the seat price. | Windsurf Teams $40 a seat a month → Teams $80 a month plus $40 per full seat; Pro $20 and Max $200 unchanged | seat | Increase | Cognition, Wayback Machine (devin.ai/pricing, 6 Jun 2026) | |
| Cursor Cursor Teams | Cursor split Teams usage into first-party and third-party pools and added a Premium seat with 5x the Standard seat's usage, for renewals from 1 July 2026. | Standard seat $40 a month → Standard seat $40 a month; Premium seat $120 a month | credits | New plan | Cursor, Cursor Docs | |
| GitHub Copilot GitHub Copilot (all plans) | GitHub replaced premium requests with GitHub AI Credits metered by tokens at API rates, at unchanged seat prices (announced 27 April 2026). Annual Pro and Pro+ subscribers stay on premium requests until renewal. | Premium request allowances; $0.04 per extra premium request → Monthly AI Credits (1 credit = $0.01): Pro $10 with $10 of credits, Pro+ $39, Business $19 a user, Enterprise $39 a user | requests to credits | Restructure | GitHub Blog | |
| Lovable Lovable | Lovable merged its separate dollar balances for Cloud and AI into the single credit balance used for building. | Build credits plus separate dollar balances for Cloud and AI → One credit balance for build, chat, Cloud, AI gateway and connectors | credits | Restructure | Lovable Docs | |
| GitHub Copilot GitHub Copilot Max | GitHub added Copilot Max, its highest individual plan, alongside the move to AI Credits. It first appears in the Copilot plans documentation on 29 May 2026. | Top individual plan: Pro+ at $39 a month → Max at $100 a month (20,000 AI credits including flex, as of October 2026) | credits | New plan | GitHub Docs history, GitHub Docs | |
| Devin Desktop (formerly Windsurf) and Devin Windsurf Pro, Teams and Max | Windsurf replaced monthly credits with daily and weekly quotas, raised Pro from $15 to $20 for new subscribers (existing prices grandfathered) and added a $200 Max plan. | Pro $15 a month with 500 monthly credits → Pro $20 with daily and weekly quotas; Teams $40 a seat; Max $200; extra usage at API pricing | credits to quota | Increase | Cognition | |
| Replit Agent Replit Core and Pro | Replit cut Core from $25 to $20 a month, added a $100 Pro plan and retired the Teams plan. | Core $25 a month; Teams plan → Core $20 a month with $20 of usage; Pro $100 a month | credits | Decrease | Replit | |
| Claude Code Claude Team | Anthropic repriced Claude Team: Standard seats got cheaper and gained Claude Code, Premium seats dropped from $150 to $100 a month billed annually with 5x a Standard seat's usage. | Standard $25 a seat annual / $30 monthly (no Claude Code); Premium $150 a seat with Claude Code, minimum 5 → Standard $20 annual / $25 monthly with Claude Code; Premium $100 annual / $125 monthly | seat | Decrease | Wayback Machine (claude.com/pricing, 26 Jan 2026), Wayback Machine (claude.com/pricing, 29 Jan 2026) | |
| Kiro Kiro | Kiro replaced separate vibe and spec request limits with one credit pool charged fractionally by prompt complexity (announced 15 September 2025, billed from 1 October). | Separate vibe and spec request limits → Free 50, Pro 1,000, Pro+ 2,000, Power 10,000 credits; overage $0.04 a credit | requests to credits | Restructure | Kiro | |
| JetBrains AI (Junie) JetBrains AI | JetBrains replaced its abstract AI quota with AI Credits worth $1 each, sized to the subscription price, with paid top-ups. | An AI quota per plan, no published unit value → AI Credits, 1 credit = $1 (AI Pro 10, AI Ultimate 35 every 30 days); top-ups valid 12 months | quota to credits | Restructure | Wayback Machine (JetBrains docs, 3 Aug 2025), Wayback Machine (JetBrains docs, 18 Sep 2025) | |
| Claude Code Claude Pro and Max (Claude Code) | Anthropic added weekly limits on top of the 5-hour limits for Pro and Max subscribers (announced 28 July 2025), at the same prices. | Rolling 5-hour limits only → 5-hour limits plus an overall weekly limit and an Opus weekly limit | quota | Increase | TechCrunch | |
| Kiro Kiro | Kiro started charging, with Free, Pro, Pro+ and Power tiers metered in separate vibe and spec requests. A metering bug led Kiro to refund August charges (22 August 2025). | n/a → Free tier with 50 vibe requests a month; paid tiers with vibe and spec request limits | requests | New plan | Kiro, Kiro | |
| Cursor Cursor Teams and Auto | Cursor moved Teams agent usage to variable request costs and ended unlimited Auto for individuals, counting Auto against included usage from renewals after 15 September 2025. | Teams: fixed request costs; Auto unlimited for individuals → Teams: variable request costs; Auto draws on included usage at token rates | requests to credits | Increase | Cursor | |
| Bolt.new Bolt.new | Bolt let unused paid-plan tokens roll over for one extra month. | Unused monthly tokens expired → Unused tokens valid for two months on an active paid plan | credits | Decrease | Bolt pricing FAQ | |
| GitHub Copilot GitHub Copilot (paid plans) | GitHub began enforcing monthly premium-request allowances, billing extra premium requests one by one. | Premium request allowances announced but not enforced → Pro 300, Pro+ 1,500, Business 300, Enterprise 1,000 premium requests a month; extra at $0.04 each | seat to requests | Increase | GitHub Changelog, The Register | |
| Replit Agent Replit Agent | Replit switched Agent from a flat fee per checkpoint to effort-based pricing, one checkpoint per request priced by the work done. | $0.25 per checkpoint → One checkpoint per request, priced by effort (below or above $0.25) | requests to credits | Restructure | Replit | |
| Cursor Cursor Pro | Cursor replaced Pro's 500 requests a month with a pool of model usage billed at API prices. After a backlash it apologised on 4 July 2025 and refunded surprise charges. | 500 requests a month for $20 → $20 of model usage at API prices a month for $20 | requests to credits | Restructure | Cursor, Cursor | |
| Cursor Cursor Ultra | Cursor added a $200 Ultra plan with 20 times Pro usage. | n/a → $200 a month, 20x Pro usage | credits | New plan | Cursor | |
| Claude Code Claude Max | Anthropic added Max plans above the $20 Pro plan. | Top individual plan: Pro at $20 a month → Max at $100 (5x Pro usage) or $200 (20x Pro usage) a month | quota | New plan | CNBC | |
| GitHub Copilot GitHub Copilot Pro+ | GitHub added a higher individual tier, Copilot Pro+, with 1,500 premium requests a month. | n/a → $39 a month with 1,500 premium requests | requests | New plan | GitHub Changelog | |
| Devin Desktop (formerly Windsurf) and Devin Devin | Devin 2.0 replaced the $500-a-month team entry price with a $20 Core plan and pay-as-you-go Agent Compute Units. | $500 a month (ACUs at $2) → Core from $20 (about 9 ACUs), then $2.25 per ACU | credits | Decrease | TechCrunch | |
| Devin Desktop (formerly Windsurf) and Devin Windsurf Pro and Teams | Windsurf stopped charging flow action credits for background tool calls, so only prompts used credits, and cut the Teams price. | Prompt credits plus flow action credits; Teams $35 a user → Prompt credits only (Pro stays $15); Teams $30 a user | credits | Decrease | TechCrunch |
26 changes, newest first. Dates are the effective date where the vendor gave one, otherwise the announcement date. "By" a date means the vendor published no date and the change is first visible in an archived capture of its pricing page on that day.
What the changes have in common
Three patterns hold across the log as of 4 October 2026. Each is a reading of the dated rows above, so check the rows before you quote the reading.
Request counts became metered pools. Cursor replaced 500 requests with $20 of usage at API prices on 16 June 2025. GitHub started billing premium requests at $0.04 each two days later, then replaced requests with token-metered AI Credits on 1 June 2026. Kiro went from separate vibe and spec request limits to one credit pool on 1 October 2025, and JetBrains put a $1 value on each AI Credit around September 2025. A request count hides how big each request is; a credit or dollar pool passes more of the model's real cost through.
Windsurf went the other way. On 19 March 2026 it dropped credits for daily and weekly quotas and raised Pro from $15 to $20 for new subscribers. Claude Code, Codex and Google's Jules also sell quotas rather than pools. A quota is easier to understand and harder to budget past, so expect the two approaches to keep trading places.
The top tier settled at $200. Claude Max 20x launched at $200 in April 2025 and Cursor Ultra followed in June 2025; Kiro Power and Windsurf Max sit at $200 too, OpenAI's Pro tiers run from $100 to $500, and Google's top Ultra tier is $199.99. Some prices also went down at the top: Google cut Ultra from $249.99 to a $99.99 tier and a $199.99 tier in May or June 2026, and Anthropic cut Claude Team Premium seats from $150 to $100 in late January 2026. For the same story across agent products outside coding, see the AI agent pricing change log.
What the pricing pages show between announcements
Not every change gets a blog post. Every week I read each tool's pricing page and note the set of prices it prints; monthly Wayback Machine captures back to January 2025 fill in the past. This table is a signal, not a log: a page that adds an annual toggle counts as a movement, and pages that only render prices in the browser can read as empty in the archive.
| Tool | Pages watched | Last read | Readable captures | Price-set movements | In the last 12 months | Last movement |
|---|---|---|---|---|---|---|
| Claude Code | 1 | 4 October 2026: 35 prices read | 15 since September 2025 | 6 | 6 | October 2026 |
| Cursor | 2 | 4 October 2026: 4 prices read | 21 since January 2025 | 5 | 3 | July 2026 |
| GitHub Copilot | 2 | 4 October 2026: 14 prices read | 22 since January 2025 | 8 | 4 | September 2026 |
| Devin Desktop (formerly Windsurf) and Devin | 1 | 4 October 2026: 12 prices read | 23 since January 2025 | 3 | 2 | June 2026 |
| OpenAI Codex | 1 | 4 October 2026: 13 prices read | 8 since January 2026 | 4 | 4 | July 2026 |
| Google Jules, Antigravity and Gemini Code Assist | 3 | 4 October 2026: 10 prices read | 15 since May 2025 | 5 | 5 | October 2026 |
| Amazon Q Developer | 1 | 4 October 2026: 6 prices read | 23 since January 2025 | 1 | 0 | August 2025 |
| Kiro | 1 | 4 October 2026: 11 prices read | 17 since July 2025 | 5 | 2 | July 2026 |
| JetBrains AI (Junie) | 1 | 4 October 2026: 9 prices read | 14 since September 2025 | 0 | 0 | none seen |
| Replit Agent | 1 | 4 October 2026: 4 prices read | 23 since January 2025 | 12 | 6 | October 2026 |
| Lovable | 1 | 4 October 2026: 6 prices read | 17 since January 2025 | 4 | 0 | September 2025 |
| Bolt.new | 1 | 4 October 2026: 6 prices read | 9 since March 2026 | 1 | 1 | September 2026 |
| v0 | 1 | 4 October 2026: 22 prices read | 16 since August 2025 | 10 | 9 | October 2026 |
| Augment Code | 1 | 4 October 2026: 6 prices read | 22 since January 2025 | 7 | 2 | July 2026 |
| Cline | 1 | 4 October 2026: the page prints no prices | 7 since September 2025 | 1 | 1 | December 2025 |
| Tabnine | 1 | 4 October 2026: the page prints no prices | 0 | 0 | 0 | none seen |
20 pricing pages read every week, plus monthly Wayback Machine captures of each tool's main pricing page. A movement is any change in the set of prices printed on the page, so it includes display changes such as a new annual toggle; 0 weekly page diffs were flagged in the last 90 days. Only confirmed moves enter the change log above.
Update log
What moved on this page, by date. Corrections are logged here rather than edited in silently.
- : Tracker launched: 16 AI coding tools, 26 verified pricing changes since April 2025, and a weekly watch of 20 pricing pages.
Methodology
What counts. Paid and free plans of AI coding tools that write or change code for you: AI editors, IDE extensions, terminal agents, cloud coding agents and browser app builders. Model API token prices are out of scope except where a tool bills them directly; they have their own tracker in the LLM API price history.
Prices. US list prices before tax, read from each vendor's pricing page or pricing documentation on the "Checked" date. Monthly billing unless marked; annual prices are shown per month where the vendor publishes them. "Entry price" is the cheapest self-serve paid plan open to new customers; region-only plans (Cursor Start in India) and closed plans (v0 Premium) are excluded. Units stay attached to every price: a seat price, a whole-team price and a one-person plan are labelled and never averaged without saying so.
Limits. Classified by what runs out: credits, tokens or a dollar pool; a rolling time-window quota; a request count; or pay per token with your own key. Where a vendor states the dollar value of included usage, or sells overage at a fixed unit price, that value is used; otherwise the field is left blank rather than estimated.
Changes. Each change cites the vendor's announcement, documentation or pricing page, or credible press. Where a vendor published no date, the change is dated by the first Wayback Machine capture that shows it and marked "by" that date; the capture before it is in the CSV. Display-only changes and temporary promotions do not count.
The watch. A scheduled job reads 20 pricing pages weekly (plain fetch, then a headless browser for pages that render prices with JavaScript) and stores the price tokens and price lines of each read. A changed price set is flagged for review; it enters the change log only once confirmed.
Known gaps. Tabnine no longer publishes prices. Enterprise contracts are not public. Google prices vary by country; the table shows the US figures. Cognition's team plan change is dated to the Devin Desktop launch on 2 June 2026; the first archived capture of the merged price list is 6 June 2026.
FAQ
How much does Claude Code cost?
Claude Code comes with a Claude subscription: Pro at $20 a month ($17 a month billed annually), Max at $100 (5x Pro usage) or $200 (20x Pro usage), and Team seats at $25 a month ($20 billed annually) or $125 for a Premium seat ($100 billed annually). Usage sits behind a rolling 5-hour limit plus weekly limits; past them you can pay for extra usage at API rates. Through the API it is billed per token instead. Prices as of 4 October 2026.
How much does Cursor cost?
Cursor Hobby is free. Pro is $20 a month, Pro Plus $60 and Ultra $200; each includes a monthly amount of model usage, with third-party models charged at their API prices and on-demand usage billed after that. Teams costs $40 a seat a month, or $120 for a Premium seat with 5x the agent limits. In India a Start plan costs INR 649 a month. Prices as of 4 October 2026.
How much does GitHub Copilot cost?
Copilot Free costs nothing. Pro is $10 a month with 1,500 GitHub AI Credits, Pro+ $39 with 7,000 and Max $100 with 20,000 (1 credit = $0.01, credits counted by tokens since 1 June 2026). Business is $19 and Enterprise $39 a seat a month. Past the allowance you set a dollar budget and keep going at $0.01 a credit. Prices as of 4 October 2026.
What is the cheapest AI coding assistant?
Among paid plans, ChatGPT Go at $8 a month, which includes OpenAI Codex for lightweight tasks, followed by GitHub Copilot Pro and JetBrains AI Pro at $10 a month, as of 4 October 2026. Free options exist too: Copilot Free, Gemini Code Assist for individuals, Cursor Hobby, and Cline, which is free software where you pay only for model tokens through your own API key.
Is Claude Code cheaper than Cursor?
They start at the same $20 a month, but they meter differently. Claude Code on Pro gives a rolling 5-hour usage window and a weekly limit shared with Claude chat; Cursor Pro gives a monthly pool of model usage charged at API prices. Claude's limit resets every 5 hours and every week, so a heavy day can stall you until the window resets; Cursor's pool is monthly, so a heavy week can use up the month and move you onto on-demand billing. Anthropic's own documentation puts average enterprise Claude Code spend at about $13 per developer per active day.
How often is this tracker updated?
Weekly. A scheduled job reads 20 pricing pages for 16 tools every week and flags any change in the prices they print; each confirmed change enters the change log with its date and source. Monthly Wayback Machine captures back to January 2025 fill in what happened between announcements. Each update is listed in the update log.
Can I reuse this data?
Yes. The plans table and the change log are published under CC BY 4.0 as CSV files on this page. Reuse them with a link back to this page. If a change is missing, send it through contact with a dated source.
Sources
- Claude pricing, Anthropic, checked 4 October 2026; Claude Code costs, Anthropic docs, checked 4 October 2026.
- Cursor pricing documentation, Cursor, checked 4 October 2026; June 2025 pricing update and refunds, Cursor, 4 July 2025.
- Plans for GitHub Copilot, GitHub Docs, checked 4 October 2026; GitHub Copilot is moving to usage-based billing, GitHub, 27 April 2026.
- Introducing our new Windsurf pricing plans, Cognition, 18 March 2026; Windsurf is now Devin Desktop, Cognition, 2 June 2026.
- Codex pricing, OpenAI, checked 4 October 2026.
- Announcing new pricing plans and Auto, Kiro, 15 September 2025; Introducing Kiro Pro Max, Kiro, 11 June 2026.
- Google AI plans, Google, checked 4 October 2026; Jules usage limits, Google, checked 4 October 2026.
- JetBrains AI licensing and subscriptions, JetBrains, checked 4 October 2026.
- Tricentis acquires Tabnine, ChannelLife, 31 July 2026.
- Every other row links its own source in the tables above; full URLs are in the two CSV files.
- My weekly pricing-page reads and monthly Wayback Machine captures, January 2025 to October 2026.
More on what AI costs: the AI agent pricing index, free LLM API tiers, the AI API reliability report and AI agent pricing explained. If you are deciding between a coding assistant and an agent that works on its own, AI agent vs copilot draws the line, and what OpenAI Codex means for agent builders covers the Codex side. Our own plans are on the pricing page.