AI Coding Tool Pricing 2026
Claude $20 / $100 / $200, Cursor from $8
As of 2026-11-18 (subscription monthly tiers last checked on 2026-11-06), three key updates have reshaped metered API and subscription allowances: on 2026-11-17 OpenAI released flagship gpt-6.2-astra (official OpenAI API list price $8 input, $0.40 cached input and $40 output per 1M tokens, down 20% from gpt-6-astra's $10/$50) and everyday go-to gpt-6.2-sol (official OpenAI API list price $2 input, $0.10 cached input and $10 output for <=272K context); on 2026-11-16 Alibaba Cloud launched pay-as-you-go qwen4-max (official Alibaba Cloud Model Studio API list price $1.50 input / $4.50 output international or ¥10 input / ¥30 output in mainland China Beijing, flat across 0 < Token <= 1M); and OpenAI Help Center (20001625) announced that starting in December 2026 Plus, Pro and Business usage allowances will be shared between ChatGPT and Codex. As of 2026-11-06, tier names and monthly prices on the official pricing pages: Claude Free $0, Pro $20 (monthly billing), Max $100, Ultra $200; The X Subscription (Grok, Cursor and X in one subscription; same prices on cursor.com/pricing) Plus $8, Premium $30, Super $100, Ultra $200; Google AI Plus $4.99, Pro $19.99, Ultra from $99.99; GitHub Copilot Pro $10, Pro+ $39, Max $100; ChatGPT tiers are Plus, Pro 100, Pro 200 and Pro 500. On 2026-10-30 Anthropic renamed Max 5x to Max and Max 20x to Ultra; prices did not change. Windsurf/Devin Pro $20 and Antigravity individual $0 are the 2026-10-20 values.
Updated 2026-11-18
One key across Claude, GPT and Chinese models, billed per token.
- Billed per token — live rates on /models
- Pay by card (Visa / Mastercard), Apple Pay, Google Pay or crypto
- Self-service: top up and activate instantly
Price map (2026-08 snapshot, rechecked 2026-11-06)
Cursor / The X Subscription paid plans from
2026-08: Cursor Pro, Claude Pro (with Claude Code) and Windsurf Pro were all $20/mo; Google AI Pro was $19.99. Rechecked 2026-10-20: Cursor is Plus $8, Premium $30, Super $100, Ultra $200. Rechecked 2026-11-06: cursor.com/pricing and the xAI docs list the same monthly prices for The X Subscription (Plus $8, Premium $30, Super $100, Ultra $200); included usage is a base amount on Plus, and 4x, 15x and 50x Plus on Premium, Super and Ultra. Cursor's page says Plus includes Grok models only and Premium and above include no allowance for third-party models, which are billed at API list price; existing Cursor plans continue, and switching is optional.
Claude Pro (Claude Code included)
Claude's official pricing page, 2026-11-06: Free $0; Pro is $20/mo billed monthly ($17 with the annual discount); Max $100; Ultra $200. Claude Code is included in all paid plans, not in Free. Max gives 5x Pro usage every 5 hours and every week; Ultra gives 10x Pro usage every week, with no 5-hour limit for verified accounts. Verification is automatic; until then a temporary limit applies, and no Current session bar in Settings > Usage means you are verified.
Windsurf / Devin Pro (current)
Checked 2026-10-20 (not refetched this time): the current Windsurf/Devin Pro tier is $20/mo, Max $200/mo, Teams $80/mo + $40/mo per seat.
Antigravity individual
$0/mo individual tier with Tab completion (checked 2026-10-20); Google AI Pro / Ultra are separate, and API use is consumption-based. Google's subscription page, 2026-11-06: AI Plus $4.99/mo, AI Pro $19.99/mo, AI Ultra from $99.99/mo (with a $199.99/mo option); AI Pro includes entry rate limits to the agent model in Antigravity.
How the convergence happened
Checked vendor by vendor on 2026-11-06: Claude Free $0, Pro $20, Max $100, Ultra $200; The X Subscription (same prices on Cursor and in the xAI docs) Plus $8, Premium $30, Super $100, Ultra $200; Google AI Plus $4.99, Pro $19.99, Ultra from $99.99 (including a $199.99 option); GitHub Copilot Pro $10, Pro+ $39, Max $100, with allowances counted in AI Credits; ChatGPT tiers are Plus, Pro 100, Pro 200 and Pro 500, with dollar prices on chatgpt.com/pricing; MiniMax M Plan has Go, Explore and Build at 1×, 3× and 7.5× the Go usage quota, and the M Plan Monthly offer page says the promotion ended on 2026-10-14 ('This offer ended on October 14, 2026.'), and its Regular price column lists regular monthly prices of $22, $55 and $132.
Earlier checks: the 2026-08 snapshot had four entry tiers at $20/mo and three heavy tiers at $100/mo (Windsurf only caught up from $15 in May); rechecked on 2026-10-20, prices have split: Cursor paid plans start at $8/mo (Plus $8, Premium $30, Super $100, Ultra $200), Claude Pro is $20/mo with Max from $100, Windsurf/Devin Pro is $20 with Max at $200, and Antigravity's individual tier is $0.
Each price comes from the official pricing page, fetched 2026-10-20 and 2026-11-06. Price is no longer the only variable; quotas, model lists and agent experience matter too.
What changed in November 2026, and the $200 tiers side by side
Mid-November updates as of 2026-11-18
- 1OpenAI's Help Center (20001625) announced that starting in December 2026, ChatGPT Plus, Pro and Business usage allowances will be shared between ChatGPT and Codex (Instant chat does not draw from that pool), and once exhausted users must wait for reset, upgrade or buy extra credits; on 2026-11-17 OpenAI launched
gpt-6.2-astra(official OpenAI API list price $8/$40) andgpt-6.2-sol(official OpenAI API list price $2/$10), opening Ultrafast speed to Pro 200 for the first time (starting withgpt-6.2-sol, while on the API Ultrafast is priced at 3x Standard through 2027-02-28). - 2On 2026-11-17 Anthropic updated its Help Center (17154008) to specify that active Claude Max ($100/mo), Ultra ($200/mo) and Team plans include monthly Claude Platform API credits (claimable after 7 days, non-rollover).
- 3On 2026-11-16 Alibaba Cloud launched pay-as-you-go
qwen4-max(official Alibaba Cloud API list price $1.50/$4.50 international or ¥10/¥30 mainland China, scoring 52.69 on Work360 at a $1.67 average task cost). $200 tiers side by side (2026-11-06; each vendor's own wording on usage, listed without ranking): Claude Ultra, $200/mo, '10x Pro usage every week' and 'No 5-hour limit' (verified accounts only); ChatGPT Pro 200, where the help center says Pro 100, Pro 200 and Pro 500 currently have no five-hour usage limit in Work and Codex but still include a usage allowance, and that the allowance is not a fixed number of messages or tasks; Google AI Ultra has a $199.99/mo option. That option's wording: '20x higher usage limits vs. AI Pro', and its Antigravity line gives it the highest agent-model rate limits among Google's plans. The X Subscription Ultra, $200/mo (same price on Cursor's page), '50x more usage compared to Plus'; third-party models in Cursor are not part of that included usage. The multipliers use different baselines (Pro for Claude, AI Pro for Google, Plus for The X Subscription) and cannot be converted into one another. On 2026-10-29 OpenAI canceled the Pro 200 reduction announced in its September 29 email, and Pro 100 and Pro 500 now include twice their previous Work and Codex allowance. Earlier, the 2026-10-20 check: Cursor now lists Plus $8, Premium $30, Super $100 and Ultra $200; Claude Pro is $20 per month with Max from $100; Windsurf's pricing page now presents the Devin brand, with Pro $20 and Max $200; Antigravity's individual tier is $0, and enterprise Standard/Plus starts at $30 per seat per month. Earlier still, on 07-30, OpenAI cut Terra by 20%; that item is kept as history. On the other side, Anthropic briefly planned (in June) to move Claude Code's programmatic calls to API billing — paused, but the direction is clear: the boundary between subscription and metered is blurring. All prices are official, checked 2026-10-20 and 2026-11-06.
Timeline
Windsurf Pro rises from $15 to $20, completing the entry-tier alignment.
OpenAI cuts prices: Terra -20%, Sol unchanged; Sol gains a 2.5x Fast mode (at 2x the price).
The two-layer structure — four at $20, three at $100 — stabilizes; pay-per-use vs subscription becomes the main selection axis.
Confirmed vs caveats
Confirmed
The 2026-11-06 prices and usage wording come from the official Claude, Cursor, xAI docs, Google subscription, GitHub Docs, MiniMax docs and ChatGPT pricing pages; the ChatGPT Work and Codex lines come from OpenAI help-center pages read on 2026-11-06. The 2026-08 snapshot came from each vendor's official pricing page or announcement; the 2026-10-20 prices come from the official Cursor, Claude, Windsurf/Devin and Antigravity pricing pages, each marked with its fetch date.
Caveats
'$20 for unlimited use' is a misconception: every $20 tier has rate and quota limits, and heavy agent usage easily hits the ceiling — which is exactly why the $100 tier exists. When comparing, look at the quota terms, not just the monthly fee.
Subscription vs pay-per-use
Subscription ($20/$100 tiers)
Predictable budget, zero hassle. Suits people who use a single tool heavily and steadily every day. Usage beyond quota is billed extra or throttled.
Pay-per-use (API billing)
Pay for what you burn, switch models freely. Suits people with fluctuating usage, multi-tool setups, or a need to mix several model families. As of 2026-11-18, official vendor API list prices per 1M tokens form a clear ladder: for the hardest reasoning tasks, OpenAI gpt-6.2-astra (official OpenAI API list price $8 input / $40 output; requests with >272K input tokens bill the entire request at 2x input $16 and 1.5x output $60) and Anthropic claude-fable-5-2 (official Anthropic API list price $8/$40); for everyday coding, OpenAI gpt-6.2-sol (official OpenAI API list price $2/$10, within 0.5 points of gpt-6.2-astra on FrontierCode 2.0 at max) and Anthropic claude-opus-5-5 (official Anthropic API list price $4/$20); and for 1M flat-rate Chinese flagships, Alibaba Cloud qwen4-max (official Alibaba Cloud API list price $1.50/$4.50 international or ¥10/¥30 mainland China, with no long-context surcharge).
How to run your own numbers
First estimate monthly tokens: light use (a few chat-level calls a day) often burns under $5 a month, below the $20 tier's monthly fee; moderate use (a few hours of pair programming daily) lands around $20 a month; only heavy use (all-day multi-agent) justifies $100. Most people's mistake is assuming they're heavy users.
On QCode
QCode is the base layer of the per-token-billing route: one key reaches several model families, billed per token, with each model’s unit price on /models. Pair it with Cline / OpenCode / Claude Code and send chores to cheap small models while flagships handle the hard tasks — the cost stays under your control.
As of 2026-11-18, QCode supports calling gpt-6.2-astra, gpt-6.2-sol and qwen4-max alongside claude-fable-5-2, claude-opus-5-5, kimi-k3.1-pro, deepseek-v4.1-flash and glm-5.3 (use https://api.qcode.cc/api for Claude Code, https://api.qcode.cc/openai for Codex CLI, and https://api.qcode.cc/openai/v1 for the OpenAI SDK, Cline and OpenCode).
Not ready for a monthly plan? Pay-as-you-go Packs start at $4 for a one-time quota you use within 7 days; if you already have a Personal Plan, add a booster pack when you need more quota.
FAQ
What do AI coding tools cost in 2026?
As of 2026-11-06, official monthly prices: Claude Free $0, Pro $20, Max $100, Ultra $200; The X Subscription (Cursor) Plus $8, Premium $30, Super $100, Ultra $200; Google AI Plus $4.99, Pro $19.99, Ultra from $99.99; GitHub Copilot Pro $10, Pro+ $39, Max $100; ChatGPT tiers are Plus, Pro 100, Pro 200 and Pro 500 (dollar prices on chatgpt.com/pricing). 2026-08 snapshot: Cursor Pro $20, Claude Pro $20, Google AI Pro $19.99, Windsurf Pro $20. 2026-10-20: Windsurf/Devin Pro $20, Max $200; Antigravity individual $0.
Is the $20 tier enough?
Depends on usage. Conversational coding is fine; all-day parallel agents are not — every tier has rate and quota terms, and heavy usage gets pushed toward the $100 tier.
What does pay-per-use cost per month?
Light users often stay under $5; moderate users land around $20; only all-day heavy agents exceed $100. Track your actual usage for a week, then estimate the monthly total.
How did per-token costs change after the GPT-5.6 price cut?
On OpenAI's price list Terra dropped to $2/$12 (-20%). Moving chores to cheaper tiers or Chinese flash models can cut the overall cost further. As of 2026-11-18, OpenAI's flagship gpt-6.2-astra released on 2026-11-17 also cut official OpenAI API list pricing by 20% compared with gpt-6-astra ($10/$50) down to $8/$40 ($0.40 cached input), while gpt-6.2-sol keeps the official OpenAI API list price of $2/$10 ($0.10 cached input); Alibaba Cloud's qwen4-max launched on 2026-11-16 has an official Alibaba Cloud API list price of $1.50/$4.50 international (¥10/¥30 mainland China, $0.075 / ¥0.5 cache hit) flat across the full 1M context window. All three models are supported on QCode, with live per-model rates on /models.
Can subscription and pay-per-use be mixed?
Yes — and it's common: the subscription tool covers the main IDE experience, while pay-per-use APIs cover automation scripts, CI and the second tool. They don't conflict.
How does QCode bill?
Billed per token; per-model rates are listed on /models, with usage and spend visible live in the console. There are two options: Personal Plans are bought by the month, and Pay-as-you-go Packs need no monthly plan — you buy one lump sum to use within 7 days.
Sources
Official pages: Claude (claude.com/pricing and the help center), Cursor (cursor.com/pricing), the xAI docs (docs.x.ai/x-subscription/pricing), Google (gemini.google/subscriptions), the GitHub Docs Copilot plans page, the MiniMax docs M Plan pages and ChatGPT (chatgpt.com/pricing), all fetched 2026-11-06; Windsurf/Devin (windsurf.com/pricing) and Antigravity (antigravity.google/pricing) were fetched 2026-10-20. Re-checked on 2026-11-18: OpenAI GPT-6.2 Astra and GPT-6.2 Sol official model docs and launch blog (2026-11-17), OpenAI Help Center shared ChatGPT and Codex usage limits article (20001625), Alibaba Cloud Model Studio qwen4-max official docs and pricing page (2026-11-16), and Anthropic Help Center monthly API credits article (17154008, 2026-11-17). The 2026-08 snapshot and the 07-30 OpenAI price-cut item remain in the body as history, not as current prices.
Estimate your monthly spend from real usage
QCode is billed per token; per-model rates are listed on /models. Personal Plans are bought by the month and Pay-as-you-go Packs need no monthly plan, so real usage answers 'should I buy the $20 tier.'
Related reading
QCode pricing explained
Billing and plan structure.
Claude Max alternative: three paths and the metered math
Checked 2026-11-04: the three Claude Max alternatives, official Pro / Max / Ultra prices and limits, metered API billing and concurrency, switching and mixing.
Qwen4-Max Official API Pricing, 2.4T MoE Specs & QCode Guide (2026-11)
Complete guide to Alibaba Cloud's 2026-11-16 GA qwen4-max: 2.4T/71B MoE + 328B N-gram Embedding, flat 1M pricing (¥10/¥30, $1.50/$4.50, $0.075 cache), low/high/xhigh thinking, Work360 benchmarks, and QCode setup.
Crypto payment guide
Nine coins, 56 networks, zero platform fee for AI coding API credits.
Prices are official-price snapshots for 2026-08, 2026-10-20 and 2026-11-06; each vendor's official pricing page is authoritative after any change. Not affiliated with the listed vendors.