Taimi — The AI Market Intelligence | Agentic Coding Tool Pricing | Kanto Company

⚠ Usage-based units differ across vendors — not directly comparable. Tokens (per 1M), credits (vendor-defined billing units), requests (per call), and ACUs (normalized agent-compute units) measure fundamentally different things. Credit-based plans add a second layer: the per-credit price is usually fixed, but how many credits a task consumes varies by vendor, model, and task complexity — and consumption rates and included allowances can change without any listed price changing. Included credits typically don't roll over; purchased top-ups often expire. Click any price block to visit official pricing.

Platform plans^P: Claude Code, OpenAI Codex, Mistral Vibe, and Devin subscription prices include their full platforms (Claude.ai, ChatGPT, Le Chat, the Devin platform) — not just the coding tool. API/usage plans are standalone.

← scroll horizontally →

Vendor ⇅

Free Tier

Individual ⇅

Team / Seat ⇅

Usage-Based (PAYG)

Enterprise

🇺🇸Augment Code

Augment Code

↗ official pricing

—

—

Business$100/mo flat

50 seats max, $100/mo usage pooled, Cosmos, CLI, MCP & Native Tools, daemon mode, 50 concurrent sessions, standard size compute, no AI training

Standard$20/mo flat

50 seats max, $20/mo usage pooled, Cosmos, CLI, MCP & Native Tools, daemon mode, 50 concurrent sessions, standard size compute, no AI training

—

Custom credits, unlimited users, SSO/OIDC/SCIM, CMEK, ISO 42001, SIEM Integration, Data Residency Options, Granular Access Controls, Comprehensive Audit Trails, Slack integration, GitHub Multi Org, Usage Analytics, Advanced Analytics, User allowlist, volume discounts, dedicated support, custom VM sizes, unlimited concurrent sessions, multi-region compute, Cosmos, no AI training

🇺🇸Claude Code

Anthropic

↗ official pricing

Claude Code requires Pro or higher

Rate-limited; Annual billing: $17/mo

5x usage vs Pro; higher output limits; early access to advanced features; priority access during peak traffic

Max 20xP$200/mo

20x usage vs Pro; higher output limits; early access to advanced features; priority access during peak traffic

Standard SeatP$25/seat

Standard seat, 2-150 members, enterprise search, central billing, SSO, admin controls, desktop app deployment, no training on user content, 200k context window; Annual billing: $20/mo

Premium SeatP$125/seat

5x more usage than Standard; enterprise search, central billing, SSO, domain verification, admin controls, desktop app deployment, no training on user content, 200k context window; Annual billing: $100/mo

API (Sonnet 4.6)$3/$15/M

API (Opus 4.8)$5/$25/M

Legacy model; available on Max/Teams Premium plans

API (Haiku 4.5)$1/$5/M

Lightweight model; available via API

API (Managed Agents)$0.08/req

$0.08 per session-hour of active agent runtime

API (Web Search)$0.01/req

$10 per 1,000 searches ($0.01/search)

API (Code Execution)$0.05/req

$0.05 per container-hour; 50 free hours per day

API (Fable 5)$10/$50/M

Legacy model as of 2026-09-02 (superseded by Fable 5.1); listed as current model on vendor page as of 2026-08-12 (previously marked unavailable 2026-07-01)

API (Sonnet 5)$2/$10/M

Legacy model as of 2026-09-30 (superseded by Sonnet 5.5); introductory $2/$10 pricing made standard; scheduled 2026-09-01 increase to $3/$15 per 1M tokens did not occur (vendor confirmed)

API (Opus 5)$5/$25/M

Legacy model as of 2026-09-23 (superseded by Opus 5.5)

API (Fable 5.1)$10/$50/M

Current flagship model as of 2026-09-02; cache reads priced at 0.025x base input rate vs standard 0.1x for other models

API (Opus 5.5)$4/$20/M

Latest Opus model; Fast mode: 2x pricing for ~2.5x faster output; cache reads priced at 0.05x base input rate vs standard 0.1x for other models

API (Sonnet 5.5)$2/$10/M

Latest Sonnet model

Enterprise Self-ServeP$20/mo

SSO, domain capture, SCIM, audit logs, RBAC, data retention controls, Compliance API, network-level access control, IP allowlisting, spend controls, Google Docs cataloging, HIPAA-ready, Claude Security (beta), 500k context window on default model; API usage billed separately; billed annually

All self-serve features plus HIPAA-ready, custom contract terms, usage commitments, PO/MSA support, inference hooks; contact salesP Institutional pricing; student/faculty access, academic research features, academic learning mode, dedicated API credits; contact for pricingP

🇺🇸Cline

Cline Bot Inc.

↗ official pricing

VS Code extension, CLI, secure client-side architecture, BYOK, any provider, MCP Marketplace, multi-root workspaces, community support

—

—

—

JetBrains, SSO/OIDC, SCIM provisioning, SLA, dedicated support, priority support, centralized billing, RBAC, team management, authentication logs, simple config management, advanced config management (coming soon), fine-grained permissioning (coming soon), inference provider restrictions, VPC deployments, OpenTelemetry

🇺🇸Cursor

Anysphere

↗ official pricing

Limited Agent requests, limited Tab completions, access to Composer

Extended Agent limits, generous Grok & Composer limits, Grok Bot access, frontier models, MCPs, skills, hooks, cloud agents, Bugbot on usage-based billing

3× Pro limits on Agent, higher Grok Bot usage limits

20× Pro limits on Agent, highest Grok Bot usage limits; priority access to new features

Teams Standard$40/seat

Cloud agents with shared team context, Bugbot code reviews, Grok Bot access, shared chats, commands, team-wide rules/skills/automations, centralized billing, privacy mode, RBAC, SAML/OIDC SSO, team plugin marketplace, usage analytics

Teams Premium$120/seat

5x Standard limits on Agent, higher Grok Bot usage limits; includes all Individual plan features

—

Pooled usage, invoice/PO billing, SCIM seat management, AI code tracking API, audit logs, service accounts, auto-run/browser/network controls, repository/model/MCP access controls, priority support, account management, Bugbot on custom plan

🇺🇸Devin

Cognition AI

↗ official pricing

Light quota to code with agents; unlimited inline edits; unlimited Tab completions; limited model availability; 1 member

ProP$20/mo flat

Devin usage quota + Devin Desktop (formerly Windsurf IDE) quota included; full model availability (OpenAI, Claude, Gemini, SpaceXAI frontier); free SWE-2 in Devin Desktop/CLI through 2026-10-16; leading open source models; Devin Cloud (cloud agents); Slack, Linear, MCP integrations; up to 10 concurrent sessions; 1 member

MaxP$200/mo flat

Increased Devin + Devin Desktop (formerly Windsurf IDE) quotas vs Pro; all Pro features; up to 10 concurrent sessions; 1 member

TeamsP$80/mo flat

All Pro features; up to 200 users; shared collaboration; centralized billing; admin dashboard with analytics; priority support; $40/month per full developer seat; unlimited flex seats; unlimited concurrent sessions

—

All Teams features; highest priority support; dedicated deployment option, SAML/OIDC SSO, dedicated Slack Connect, early feature releases, centralized enterprise admin controls, dedicated account management, custom terms, teamspace isolation; unlimited concurrent sessionsP

🇺🇸GitHub Copilot

Microsoft / GitHub

↗ official pricing

50 agent mode/chat req/mo, 2K completions, Haiku 4.5/GPT-5 mini access

300 premium req/user; $15/mo AI Credits ($10 base + $5 flex); legacy premium req/user; unlimited completions; third-party agent access (Claude Code, Codex); Cloud agent; code review; signups gradually re-enabling (2026-06-21)

1500 premium req/user; $70/mo AI Credits ($39 base + $31 flex); legacy premium req/user; GitHub Spark (preview); third-party agent delegation; Claude Opus 4.7 access; Audit logs; signups gradually re-enabling (2026-06-21)

$200/mo AI Credits ($100 base + $100 flex); new signups paused 2026-04-20; for existing Copilot subscribers only from 2026-06-01; signups gradually re-enabling (2026-06-24)

Business$19/seat

300 premium req/user; legacy premium req/user; IP indemnity; org management; self-serve signups paused 2026-04-22

—

Enterprise$39/mo

1000 premium req/user; legacy premium req/user; all models incl Claude Opus 4.7; GitHub Spark; IP indemnity; SSO; codebase indexing; fine-tuned custom/private models

🇺🇸Google Antigravity

Google

↗ official pricing

Basic weekly rate limits; unlimited Tab completions; unlimited Command requests; Gemini 3.8 Flash, Gemini 3.7 Flash, Gemini 3.6 Flash, Gemini 3.1 Pro, Claude Sonnet & Opus 4.6, gpt-oss-120b

Google AI Pro$19.99/mo

More generous rate limits than Individual; quota refreshes every 5 hours until weekly limit; flexible AI credit pool

Google AI Ultra$99.99/mo

Higher access to latest Gemini models vs Google AI Pro; highest weekly rate limits; quota refreshes every 5 hours; flexible AI credit pool

—

—

Consumption-based API pricing via Google Cloud; Standard/Plus plans $30/seat/mo, pay-as-you-go has $0 seat fee; Google Cloud Project integration; access under Google Cloud Terms of Service; Antigravity 2.0 and CLI via Gemini Enterprise Agent Platform; included in select Gemini Enterprise app subscriptions; rolling out to select customers (2026-08-05)P

🇺🇸Kiro

Amazon Web Services

↗ official pricing

50 credits/mo; open-weight models (Qwen3 Coder Next, DeepSeek 3.2, MiniMax M2.1) and Claude Sonnet 4.5 (social login or AWS Builder ID sign-ups only); credits reset monthly, unused credits do not roll over; $20 credit on first upgrade to any paid plan (social login or AWS Builder ID only); unavailable in AWS GovCloud regions

1,000 credits/mo; Auto, Claude Sonnet 5 and Claude Opus 5; credits reset monthly; GovCloud regions: ~20% higher price

2,000 credits/mo; Auto, Claude Sonnet 5 and Claude Opus 5; credits reset monthly; GovCloud regions: ~20% higher price; Startup program: up to 1 year free (application required)

10,000 credits/mo; Auto, Claude Sonnet 5 and Claude Opus 5; credits reset monthly; GovCloud regions: ~20% higher price

5,000 credits/mo; Auto, Claude Sonnet 5 and Claude Opus 5; credits reset monthly; GovCloud regions: ~20% higher price

—

—

Centralized billing, SSO, usage analytics, enterprise security controls; billed through AWS

🇪🇺Mistral Vibe

Mistral AI

EU-based vendor

↗ official pricing

Le Chat Free — limited coding sessions, chat, search, image generation, Mistral Studio access; $10/mo API credits included; 100+ connectors

Mistral Vibe (CLI/IDE/web coding); up to 15GB document storage; $30/mo API credits included; $5.99/mo for verified students; chat and email support

TeamP$24.99/seat

Collaborative workspace, 30GB storage per user, domain verification, Admin API, data export; $50/mo minimum

Devstral 2 API$0.40/$2/M

123B param, 256K context; DEPRECATED: Mistral docs (docs.mistral.ai, 2026-08-16) list Devstral 2 as retired 2026-07-31, alternative Mistral Medium 3.5 (general-purpose, not a coding successor)

Custom pricing, private deployments, custom models/agents/workflows and UI; SAML SSO, audit logs, white label, custom SLAs; Mistral Vibe for IDE, async coding agentsP

🇺🇸OpenAI Codex

OpenAI

↗ official pricing

ChatGPT Free subscription; basic exploration of Codex capabilities; GPT-6 Luna, desktop app (subject to rollout)

ChatGPT Plus subscription; GPT-6.1 Sol, GPT-6 Luna, GPT-6 Astra, GPT-5.6 Sol/Terra/Luna, GPT-5.5 (retires 2026-10-14), GPT-5.4, GPT-5.4 mini; web, CLI, IDE extension, iOS; cloud integrations: code review, Slack; flexible credit purchases

ChatGPT Pro 5x subscription; GPT-6 Sol and Luna (same as Plus); 5x Plus rate limits; priority processing

ChatGPT Go subscription; lightweight coding tasks; may include ads; GPT-6 Luna, desktop app (subject to rollout)

Pro 20xP$200/mo

ChatGPT Pro 20x subscription; GPT-6 Sol and Luna (same as Plus); 20x Plus rate limits; priority processing; unlimited ChatGPT Voice

BusinessP$25/seat

Annual billing: $20/user/mo; minimum 2 users; desktop and mobile access; larger VMs for faster tasks; same usage limits as Plus; secure workspace; SAML SSO; MFA; no training on business data by default; credit flexibility; limited-time: up to $500 credits for new Codex teams (no published end date)

API key mode billed at standard API rates; model availability follows the API models available to your key; GPT-5.6 Sol/Terra/Luna, GPT-5.5, GPT-5.4, GPT-5.4 mini confirmed available; no cloud features

Custom pricing; priority request processing; SCIM, EKM, RBAC, domain verification, user analytics, audit logs via Compliance API, data retention and residency controlsP

🇺🇸Replit

Replit

↗ official pricing

Limited daily Agent credits, 1 published app

$20/mo credits included, full Agent access, 5 collaborators, up to 2 parallel agents

Pro$100/mo flat

$100/mo credits, up to 15 collaborators, 50 viewers, premium support, up to 10 parallel agents, database rollback up to 28 days

—

Custom pricing, dedicated support, SSO/SAML, single-tenant environments, custom seat limits, static outbound IPs

Key observations

OpenAI Codex Go ($8/mo, platform^P) is the cheapest paid subscription; GitHub Copilot Pro ($10/mo) is the cheapest standalone individual plan. Mistral Vibe ($14.99/mo) is the only EU-headquartered vendor tracked here — critical for data sovereignty.
Team seat costs range ~6.6× from $19 (Copilot Business) to $125 (Claude Code Premium Seat, monthly billing); Kiro dropped its team tiers entirely (absent from its pricing page as of 2026-08-30), leaving Enterprise as its only multi-seat option.
Usage-based is apples to oranges: Copilot moved to AI Credits (token-based, $0.01/credit) June 2026 — previously per request ($0.04). Anthropic/OpenAI/Mistral charge per token, Replit per effort-based Agent task; Devin publishes no per-unit rate. Cursor bills on-demand usage in dollars, in arrears, once the included pool is spent (Pro+/Ultra buy 3×/20× larger pools).
Among current token rates tracked here, Claude Haiku 4.5 output ($5/M) is the cheapest — 10× less than Claude Fable 5.1 output ($50/M). Claude Sonnet 5 ($10/M output) sits near the low end — the scheduled 2026-09-01 increase to $15/M was cancelled, making the introductory rate permanent. (Mistral's Devstral 2 rate of $2/M output is still listed on the Mistral plans but excluded here: the vendor retired the model on 2026-07-31.)
Platform bundling^P: Claude Code, Codex, Vibe, and Devin subscription plans include their full platforms (Claude.ai, ChatGPT, Le Chat, and — since Cognition folded Windsurf into Devin Desktop in June 2026 — the Devin Desktop IDE). Their prices aren't directly comparable to standalone tools like Cursor or Kiro. Google Antigravity (agent-first IDE, Free–$100/mo tracked) replaced Gemini Code Assist for individual developers at Google I/O 2026.
Self-hosted / on-prem: Mistral (EU data, fine-tune), Cline (open-source BYOK). Kiro (AWS) is the only AWS-native tool — spec-driven IDE with hooks and steering files; Pro at $20/user. Replaces Amazon Q Developer (IDE plugins and subscriptions EoS April 2027).
2026 shift: subscription overage increasingly runs through credit systems. GitHub Copilot replaced premium requests with token-metered AI Credits (June; $0.01/credit, billed in arrears against a budget); Kiro sells prepaid credit packs (July; $0.04/credit); Google Antigravity meters overage in purchasable credits; Claude's opt-in usage credits are prepaid and bill at standard API token rates. The per-credit price is visible — the number of credits a given task will consume is rarely predictable in advance. Two tracked vendors run the other way in plain dollars: Cursor (in arrears) and Augment, which dropped credits for list-price tokens plus a disclosed flat 40% service fee. Purchased top-ups commonly expire 12 months after purchase (Kiro, Augment).
No SWE-bench scores here, on purpose. Vendor publications are self-reported with cherry-picked configurations, the official leaderboard doesn't cover consumer agent products (it ranks model+scaffold combos), and variants (Verified, Pro, Lite) aren't comparable. Showing a single number next to "—" would imply a ranking the data can't support.