O preço de cada ferramenta de IA que monitoramos. Verificado, datado, público.

Correções que fizemos em nossos próprios dados

• cursor: Plus/Premium/Super/Ultra/Teams billingType usage→monthly (start price stays $8, not $0); Enterprise set to quote-only form
• codex: Plus/Pro/Business billingType usage→monthly; removed 'API Key (pay-as-you-go)' plan; Enterprise & Edu set to quote-only form
• replit: Accepted file: Replit Core $25→$20/mo ($216/yr), Replit Pro annual $1,140→$1,080
• command-code: Accepted Pro $15→$20 and new Max 20x $200; Provider billingType usage→monthly ($15/mo subscription + no-markup pay-as-you-go); removed unverified 'Beam 1.0' from offer
• opencode: Offer: removed unverified '30% off open-source models through ~2026-11-10', replaced with promos shown on the page; Zen note replaced with the Console's published terms; Go Teams note corrected (per-member limits, not pooled)
• warp: Limits rewritten: included usage is now stated in dollars ($20 Build, $200 Max, $20/seat Business), not credits; Free has no included amount
• zcode: Kept DB plan names 'GLM Coding Plan — Lite/Pro/Max' (file changed only the dash)
• google-jules: Kept DB plan name 'Free' (file renamed to 'Free (Jules base)'; vendor calls the base tier just 'Jules')
• xai-api: Accepted rename grok-4-7→grok-4.7
• glm-zai: 9 models: per-token input rate removed from monthlyUsd (now 0, usage; rates stay in usagePriceNote); restored 12 still-listed models the file dropped (GLM-5.2, 5.1, 4.7-FlashX, 4.5-X, 4.6V-FlashX, 4.6, 4.5, 4.5-AirX, 4-32B-0414-128K, 4.6V-Flash, 4.5V, ASR-2512); GLM-5-Turbo removed (no longer listed); cached-input 'Limited-time Free' offer kept
• minimax: Restored 'MiniMax-M3 Priority (≤512k/>512k input)' rows dropped on 'no change'; accepted M Plan annual $220/$550/$1,320
• qwen: Kept DB name 'qwen3.7-max (International/Singapore)'; removed the unsupported '50% off' claim from its note and the offer; restored qwen3.6-plus (still listed)
• kimi: Kept DB name 'K3 (flagship)' with its fuller note (cache-write rates); added cache-write rates to K3.1 Pro HighSpeed
• hermes: Accepted Free/Plus $20/Super $100/Ultra $200 ladder; corrected the DeepSeek V4 Flash sample rate ($0.06/$0.30 → $0.05/$0.16)
• groq: Kept freePlanAvailable true; gated-models tier set to quote-only form (model list moved to limits). Prices accepted.
• chatgpt: Accepted Pro split into Pro 100/200/500 ($100/$200/$500, monthly only) and Business split into Standard ($25 monthly / $20 annual) and Premium ($125 / $100); Enterprise set to quote-only form
• claude: Accepted new Ultra $200/mo (10x weekly) and Team split into Team Standard ($25/$20) and Team Premium ($125/$100); Enterprise $20/seat KEPT (not refused) — the vendor page states per-seat plus usage
• claude-code: Enterprise $20/seat KEPT (per-seat + usage exception verified); Team Standard/Premium/Enterprise billingType monthly→annual (as on the claude slug); Max note corrected to 5x only; API note rewritten with Claude Fable 5.2 $8/$40 and corrected legacy cache-hit rates
• openai-api: API note rewritten: restored still-listed models the file dropped (gpt-6-sol, gpt-5.6-terra, gpt-5.5, gpt-5.4 family), added gpt-6.1-sol Ultrafast, corrected image model names (gpt-image-2.6-sunburst/-flare)
• google-jules: Removed unsupported '4x usage of the $100 Ultra tier' from Ultra 20x limits
• square: Rejected the file's 'Free (Starter)' plan and its ⚠️ $0.10 in-person fee; kept DB plan 'Free' at 2.6% + 15¢ in person (added ACH 1%)
• attio: Plus $44→$50 and Pro $99→$125 (now per workspace incl. 1 user; per-user rates are extra-user add-ons); file's quote-only 'Enterprise' replaced by Ultra $1,000/mo billed annually
• julius: Business: kept DB 'up to 50 team members' (the '10 seats' claim is not on the page); Growth credits 120,000→100,000/mo and unsupported '30+ people' removed; Enterprise quote-only form. All prices accepted.
• apollo: ⚠️ Basic $49→$65 and Professional $79→$99 monthly-billed ($49/$79 are the annual-billed per-month rates, kept as $588/$948 per year); Organization $119→$149/seat, annual only, min 3 seats
• gamma: Prices carried forward from DB (USD not verifiable: page served INR); Plus credits 1,000/2,000→1,500/3,000; removed 'monthly-only' from Ultra; offer 28%→'up to 26%'
• otterly-ai: Added annual prices $25/$161/$416 per month ($300/$1,932/$4,992 per year); limits updated to 20/160/535 prompts; Enterprise quote-only form
• cal-com: Kept DB plan name 'Free'; Teams/Organizations monthlyUsd now the monthly-billing price $16/$37 (yearly $12/$28 kept as $144/$336 per year); Enterprise quote-only form
• n8n: monthlyUsd now the monthly-billing price (EUR24/60/800 converted ≈ $26.88/$67.19/$895.89), annualUsd the real yearly total (EUR240/600/8,000); billingType annual→monthly. Exchange rate is the cowork file's, not re-checked.
• buffer: Accepted removal of Agency plan
• stripe: Note carried forward from DB; the file's new surcharge figures were not verified
• kling: Premier kept at the $92 8,000-credit price ($728.64/yr) as its 'from' price instead of $207 (the slider's default 18K position); Ultra+ annual $4,902.48→$4,902.99. Accepted Ultra→Ultra+ ($180→$619, 57,000 credits) as a vendor rename/reprice.
• recraft: Restored the Teams plan the file dropped on 'no change', and corrected it: monthly $18→$22/seat ($18 is the annual-billed rate; $211/yr)
• lindy: Accepted Plus/Pro/Max → single Team plan, but the Team note now discloses the live A/B test ($29.99/3,000 credits vs $49.99/5,000 credits; FAQ says $49.99); Enterprise quote-only form. OWNER DECISION: hold Lindy out of this apply until the test ends if you prefer.
• vapi: Accepted Build/Scale → No Success Package / Core $29 (prepaid balance) / Pro ($999/mo minimum) / Premier; Premier set to quote-only form
• thunderphone: Accepted Spark/Bolt/Storm → cost-plus pricing; rewrote the note to match the page (17-25% margin on price, example rates 2.03¢/4.74¢/6.81¢ — the file's ~9¢/~11¢ examples are not on the page)
• cartesia: Removed annual prices $48/$468/$2,868 (the page has no annual toggle or annual prices); offer reworded
• suno: Pro $8→$10/mo and Premier $24→$30/mo; annual $72→$96 and $216→$288 (DB and file both stored the yearly-billed per-month rate as the monthly price)
• higgsfield: Accepted Starter/Plus/Ultra ($19/$59/$129) → Basic/Pro/Max ($9/$29/$79), Free plan removed
• bland: Accepted removal of Scale ($499/mo); Enterprise quote-only form
• openrouter: 'Enterprise' written in the quote-only form (monthlyUsd 0, billingType annual, 'Custom pricing — contact sales')
• zapier: 'Enterprise' written in the quote-only form (monthlyUsd 0, billingType annual, 'Custom pricing — contact sales')
• make: 'Enterprise' written in the quote-only form (monthlyUsd 0, billingType annual, 'Custom pricing — contact sales')
• elevenlabs: 'Enterprise' written in the quote-only form (monthlyUsd 0, billingType annual, 'Custom pricing — contact sales')
• retell: 'Enterprise' written in the quote-only form (monthlyUsd 0, billingType annual, 'Custom pricing — contact sales')
• smith-ai: 'Virtual Receptionist Enterprise' written in the quote-only form (monthlyUsd 0, billingType annual, 'Custom pricing — contact sales')
• heygen: 'Enterprise' written in the quote-only form (monthlyUsd 0, billingType annual, 'Custom pricing — contact sales')
• runway: 'Enterprise' written in the quote-only form (monthlyUsd 0, billingType annual, 'Custom pricing — contact sales')
• descript: 'Enterprise' written in the quote-only form (monthlyUsd 0, billingType annual, 'Custom pricing — contact sales')
• canva: 'Enterprise' written in the quote-only form (monthlyUsd 0, billingType annual, 'Custom pricing — contact sales')
• copy-ai: 'Enterprise' written in the quote-only form (monthlyUsd 0, billingType annual, 'Custom pricing — contact sales')
• clay: 'Enterprise' written in the quote-only form (monthlyUsd 0, billingType annual, 'Custom pricing — contact sales')
• vercel: 'Enterprise' written in the quote-only form (monthlyUsd 0, billingType annual, 'Custom pricing — contact sales')
• supabase: 'Enterprise' written in the quote-only form (monthlyUsd 0, billingType annual, 'Custom pricing — contact sales')
• resend: 'Enterprise' written in the quote-only form (monthlyUsd 0, billingType annual, 'Custom pricing — contact sales')

Correções que fizemos em nossos próprios dados

• codex: API Key (pay-as-you-go) plan not applied (subscription plans only)
• glm-zai: usage plans monthlyUsd normalised to 0 (per-token rate stays in usagePriceNote); existing offer kept
• openai-api: plan row left as-is (stamp only); file note drops gpt-6-sol / gpt-5.6-terra / gpt-5.5, live page lists them
• kimi: existing offer kept
• qwen: qwen3-max (tiered) removed; minimax: M2.7 and M2.7-highspeed removed, Token Plan renamed to M Plan
• cursor: applied as written (unified X Subscription tiers, owner confirmed); applied in a second pass after the other 15 tools

Correções que fizemos em nossos próprios dados

• cursor: claimed 'standalone plans replaced by X Subscription' independently CONFIRMED on cursor.com/pricing 2026-10-20 (Free/Plus 8/Premium 30/Super 100/Ultra 200/Teams 30+); applied; removed 'Cursor Pro $20' as the entry price (min now $8).
• grok-bot: Premium+/Teams bundling confirmed on cursor.com Grok Bot column (Free/Plus not included); applied via in-place renames, SuperGrok rows carried forward (x.ai/pricing returned 403 to our fetch).
• openai-api: NOT applied. Claim that gpt-6-sol/gpt-5.6-terra/gpt-5.5/gpt-5.4/gpt-5.4-mini are absent was contradicted by developers.openai.com/api/docs/pricing on 2026-10-20 (all listed). Existing DB rate card kept, last_verified_at bumped only.
• glm-zai: usage plans had per-1M token prices in monthlyUsd (fault); forced monthly to 0 for billingType usage. Only the 7 new rows inserted; existing rows untouched.
• claude-code: file note still said Haiku 4.5 $1/$5 despite summary; corrected to Haiku 5.5 $0.10/$0.50 (<=100K) / $0.50/$2.50 (>100K) per platform.claude.com pricing; Sonnet 5.5 cache read corrected $0.20 -> $0.10 per same page.
• grok-voice: $0.08/min speech-to-speech confirmed on docs.x.ai (Grok Voice 3.0 is current; 2.0 deprecated).
• claude: Enterprise $20 <= Pro $20 trips the Enterprise guard; not rewritten, only last_verified_at bumped.
• qwen/minimax: 4 and 2 new usage plans inserted respectively; existing plans unchanged. Other tools (chatgpt, gemini, perplexity, codex, deepseek, kimi, openrouter, xai-api): prices identical, last_verified_at bumped; offers set only for perplexity and replit.

Correções que fizemos em nossos próprios dados

• Superseded for overlapping slugs by 2026-10-16; only replit taken solely from this file (DB already held Core $20 / Pro annual $1,080; refreshed offer text and last_verified_at).

Correções que fizemos em nossos próprios dados

• glm-zai: set monthlyUsd to 0 on 13 usage plans (per-1M prices live in usagePriceNote; a non-zero monthly would render as a subscription price) and converted 12 literal "None" strings to null
• openai-api: kept the re-verified API plan only; the three carried-forward rows (gpt-5.6-terra, gpt-5.5, gpt-5.4 family) are folded into its usagePriceNote with an explicit not-re-verified label instead of being published as separate plans
• chatgpt: kept the previously verified Free-plan limits text (file reworded it to "Limited" without a source; chatgpt.com/pricing returned 403 to re-check)
• glm-zai, kimi: kept the existing tool-level offers; the file omitted them without saying they ended
• claude: Enterprise $20/seat kept — guard exception satisfied: claude.com/pricing (2026-09-29) says "Seat price + usage at API rates", US$20/seat/month billed annually, and the plan note states that structure

Correções que fizemos em nossos próprios dados

• bland/Build billingType monthly->usage
• bland/Scale billingType monthly->usage
• recraft/Pro annualUsd 16->192 (12 x $16)
• julius/Ultra preserved from DB ($500/mo)
• ideogram/Team preserved from DB ($30/mo)
• recraft/Teams preserved from DB ($18/mo)
• opencode/OpenCode Zen monthlyUsd null->0
• hermes/Hermes Cloud (new hosted daily-billed instance) monthlyUsd null->0