Corporate AI Coding Tools: Pricing & Feature Matrix (March 2026)
Last updated: March 16, 2026
Executive Summary
- GitHub Copilot remains the market leader with 20M+ total users, 4.7M paid subscribers, and deployment at ~90% of Fortune 100 companies, but faces intensifying competition from Cursor (~$2B ARR, $29.3B valuation) and Claude Code (fastest-growing tool in 2025-2026).
- Pricing has converged around $10-20/seat/month for professional tiers and $39-60/seat/month for enterprise, with most vendors now using credit-based systems rather than flat unlimited models. The cheapest full-featured option is Windsurf Pro at $15/month; the most expensive enterprise tier is Google Gemini Code Assist Enterprise at $54/user/month (via Google Cloud).
- Enterprise features are now table stakes: SSO/SAML, audit logs, and usage analytics are standard across all enterprise tiers. Key differentiators are IP indemnification (only GitHub Copilot and Google Gemini Code Assist explicitly offer this), self-hosted/air-gapped deployment (Tabnine and JetBrains AI Enterprise lead here), and code customization on private repositories (Gemini Code Assist Enterprise, GitHub Copilot Enterprise).
- Major market disruptions in 2025-2026: Windsurf (formerly Codeium) was acquired by Cognition AI for ~$250M after a failed $3B OpenAI acquisition attempt; Sourcegraph discontinued Cody Free and Pro to pivot toward its new “Amp” agentic tool; credit-based pricing models replaced flat-rate subscriptions at Cursor, Windsurf, and JetBrains; and Tabnine launched a two-tier platform model separating Code Assistant ($39/user/mo) from Agentic ($59/user/mo) capabilities.
- For a law firm evaluating tools, GitHub Copilot Business ($19/seat/month) offers the best combination of IP indemnification, enterprise governance, and broad IDE support. Cursor Teams ($40/seat/month) is the strongest for agentic/autonomous coding workflows. Tabnine Enterprise ($39-59/seat/month) is the only option offering true air-gapped, on-premises deployment for maximum data control.
Comparison Table
| Tool | Free Tier | Pro/Individual | Business/Teams | Enterprise | IP Indemnity | Self-Hosted | SSO/SAML |
|---|---|---|---|---|---|---|---|
| GitHub Copilot | $0 (2K completions, 50 chats) | $10/mo (Pro) / $20/mo (Pro+) | $19/user/mo | $39/user/mo | Yes (Business+) | No (GHEC only) | Yes |
| Cursor | $0 (2K completions, 50 slow requests) | $20/mo (Pro) / $60/mo (Pro+) / $200/mo (Ultra) | $40/user/mo (Teams) | Custom | No | No | Yes (Enterprise) |
| Amazon Q Developer | $0 (50 chats, 10 agents/mo) | $19/user/mo (Pro) | – (Pro covers teams) | – (Pro via IAM) | No | No | Yes (via IAM Identity Center) |
| Tabnine | $0 (basic completions) | $12/user/mo (Dev) | – | $39/user/mo (Code Assist) / $59/user/mo (Agentic) | Yes (IP indemnification) | Yes (air-gapped) | Yes |
| Sourcegraph Cody | Discontinued (Jul 2025) | $9/user/mo (Pro) – discontinued | – | $59/user/mo | Not confirmed | Yes | Yes |
| Windsurf | $0 (25 credits/mo) | $15/mo (Pro) | $30/user/mo (Teams) | $60/user/mo | No | Yes (Enterprise) | Yes (Enterprise; +$10/user on Teams) |
| JetBrains AI | $0 (3 credits/30 days) | $8/mo (Pro) / $30/mo (Ultimate) | – | Custom (AI Enterprise) | No | Yes (on-prem models) | Yes (Enterprise) |
| Google Gemini Code Assist | $0 (6K code reqs + 240 chats/day) | Included w/ Google AI Pro/Ultra | $22.80/user/mo (Standard) | $54/user/mo (Enterprise) | Yes | No (GCP only) | Yes (via Google Cloud IAM) |
Detailed Findings Per Tool
1. GitHub Copilot
Vendor: GitHub (Microsoft)
Pricing Tiers (March 2026):
| Tier | Price | Premium Requests/mo | Key Differentiators |
|---|---|---|---|
| Free | $0 | 50 | 2,000 code completions, basic chat, Copilot CLI |
| Pro | $10/mo ($100/yr) | 300 | Unlimited completions, coding agent, code review, multi-model (Anthropic, Google, OpenAI) |
| Pro+ | $20/mo | 1,500 | GitHub Spark access, enhanced model selection |
| Business | $19/user/mo | 300/user | User management, usage metrics, IP indemnity, org-wide policies |
| Enterprise | $39/user/mo | 1,000/user | All models incl. Claude Opus 4.6, knowledge bases, custom models, GitHub.com chat |
Enterprise Features:
- SSO/SAML via GitHub Enterprise Cloud
- Audit logs: seat assignments, policy changes, access revocations (180-day retention, JSON/CSV export)
- IP indemnification for unmodified suggestions (when filtering is enabled)
- SOC 2 Type II, ISO 27001 compliance
- No self-hosted Copilot option (requires GitHub Enterprise Cloud)
- SCIM provisioning for user lifecycle management
Adoption Numbers:
- 20M+ total users (July 2025); 4.7M paid subscribers (Jan 2026, ~75% YoY growth)
- 90% of Fortune 100 companies; 50,000+ organizations
- 42% market share of AI coding tools market ($7.37B in 2025)
- Copilot now generates 46% of code written by its users
Recent Changes (2025-2026):
- Introduced 5-tier pricing (added Free and Pro+ tiers)
- Switched to premium request-based model with multi-model access
- Added coding agent capabilities (autonomous PR creation, code review)
- Released GitHub Spark for rapid prototyping
- Copilot revenue now exceeds what GitHub itself generated at time of 2018 acquisition
2. Cursor
Vendor: Anysphere Inc.
Pricing Tiers (March 2026):
| Tier | Price | Credits/mo | Key Differentiators |
|---|---|---|---|
| Hobby (Free) | $0 | Limited (50 slow premium requests) | 2,000 completions, basic chat |
| Pro | $20/mo ($16/mo annual) | $20 credit pool | Unlimited Tab completions, Agent requests, premium models |
| Pro+ | $60/mo ($48/mo annual) | $60 credit pool | Higher usage limits |
| Ultra | $200/mo | $200 credit pool | Maximum usage |
| Teams | $40/user/mo | Per-user pools | Admin dashboard, centralized billing, usage analytics, SOC 2 |
Enterprise Features:
- SAML-based SSO (Okta, Entra ID, Google Workspace)
- SCIM provisioning
- SOC 2 Type 2 certified, annual penetration testing
- AES-256 encryption at rest, TLS 1.2+ in transit
- IP allowlisting, audit logs (Enterprise tier)
- No IP indemnification offered
- No self-hosted option
Adoption Numbers:
- 1M+ users; 360,000 paying customers
- 50,000+ enterprise seats across Fortune 1000; ~60% of revenue from enterprise customers
- ~$2B ARR (2026), up from $1.2B in 2025 (up 1,100% YoY from 2024)
- $29.3B valuation (Nov 2025 Series D, $2.3B raised)
- Fastest SaaS company ever from $1M to $500M ARR
- Notable enterprise customers include OpenAI, AB InBev
Recent Changes (2025-2026):
- Switched from request-based to credit-based pricing (June 2025)
- Credit pool equals plan price in dollars; depletes based on model used
- Annual billing saves 20%
- Cursor Agent now #1 agentic AI tool (19.3% adoption vs. 18% for Copilot Agent)
- Added Bugbot product for automated PR review ($40/user/mo for Pro and Teams)
- Added Ultra tier ($200/mo) for power users with 20x usage multiplier
3. Amazon Q Developer
Vendor: Amazon Web Services (AWS)
Pricing Tiers (March 2026):
| Tier | Price | Key Limits | Key Differentiators |
|---|---|---|---|
| Free | $0 | 50 chats/mo, 10 agent invocations, 1,000 lines of code transformation | Basic IDE integration, limited AWS resource queries |
| Pro | $19/user/mo | Unlimited chats, unlimited agent invocations, 4,000 lines of code transform (overage: $0.003/line) | Unlimited AWS queries, generative SQL, analytics dashboard, user management |
Enterprise Features:
- SSO via AWS IAM Identity Center (connects to existing IdP)
- Audit logging via AWS CloudTrail and CloudWatch Logs
- VPC endpoints for network isolation
- IAM-based access control and policy management
- No self-hosted option (AWS managed service)
- No separate Enterprise tier – Pro covers enterprise needs
- No explicit IP indemnification
Adoption Numbers:
- No specific user count disclosed publicly
- Thousands of developers reported using daily
- AWS reported 4,500 developer-years and $260M saved internally in 2024
- 12X adoption surge at Epsilon during 2025
- Netsmart: 3M+ AI interactions in 2025
Recent Changes (2025-2026):
- Renamed from Amazon CodeWhisperer to Amazon Q Developer (April 2024)
- Added agentic capabilities (unlimited agent invocations on Pro)
- Code transformation for Java modernization expanded
- Generative SQL capabilities added
- Deep AWS service integration (strongest for AWS-centric shops)
4. Tabnine
Vendor: Tabnine (formerly Codota)
Pricing Tiers (March 2026):
| Tier | Price | Key Differentiators |
|---|---|---|
| Basic (Free) | $0 | AI-powered code completions |
| Dev | $12/user/mo | Advanced completions, chat, agents for small teams |
| Code Assistant Platform | $39/user/mo (annual) | AI completions, chat, Jira integration, multi-LLM support |
| Agentic Platform | $59/user/mo (annual) | All Code Assistant features + autonomous agents, CLI, MCP support, unlimited codebase connections, Coaching Guidelines |
Enterprise Features:
- SSO (SAML) with major IdPs; SCIM for user lifecycle management
- Flexible deployment: SaaS, VPC, on-premises, fully air-gapped
- Zero code retention policy; no training on user code
- GDPR, SOC 2, ISO 27001 compliance
- Enterprise Context Engine for codebase-aware suggestions (launched Feb 2026)
- Strongest self-hosted/air-gapped option in the market — partnered with Dell for turnkey GPU-accelerated air-gapped deployment (showcased at NVIDIA GTC 2025)
- IP indemnification (license-safe AI)
- End-to-end encryption and TLS
- Admin Suite with data analytics, model provisioning, context permissions
Adoption Numbers:
- 1M+ active developers; 9.1M VS Code installs
- ~5% developer adoption rate (vs. 68% for Copilot among AI users)
- Enterprise clients report 30% faster code delivery
- Niche strength in regulated industries (finance, defense, healthcare)
- Won InfoWorld Technology of the Year Award 2025 for Software Development Tools
Recent Changes (2025-2026):
- Launched two-tier enterprise model: Code Assistant Platform ($39/user/mo) and Agentic Platform ($59/user/mo)
- Added Enterprise Context Engine for deep codebase integration (GA Feb 2026)
- Agentic Platform adds autonomous agents, Tabnine CLI, MCP support for Git/testing/linters/Docker/CI/CD
- Coaching Guidelines feature for organizational standards adherence
- Focused positioning on privacy-first, air-gapped enterprise use cases
- Only vendor offering fully air-gapped deployment while competitors (Windsurf, Cursor, Copilot) remain cloud-only or have withdrawn on-prem options
5. Sourcegraph Cody
Vendor: Sourcegraph
Pricing Tiers (March 2026):
| Tier | Price | Status | Key Differentiators |
|---|---|---|---|
| Free | $0 | Discontinued (Jul 2025) | Was: unlimited completions, 200 chats/mo |
| Pro | $9/user/mo | Discontinued (Jun 2025) | Was: 1M token context, priority features |
| Enterprise | $59/user/mo | Active | AI + code search, enterprise security, scalability |
Enterprise Features:
- Enterprise-grade security and compliance
- Deep code search across entire codebase (Sourcegraph’s core strength)
- Scalable to large monorepos
- Self-hosted deployment options
- SSO and access controls
Adoption Numbers:
- Enterprise customer base not publicly disclosed post-pivot
- Sourcegraph is transitioning users to “Amp” (new agentic tool)
- $40 in Amp credits offered to transitioning Cody Pro/Enterprise Starter users
Recent Changes (2025-2026):
- Major pivot: Discontinued Cody Free and Pro plans (Jun-Jul 2025)
- Introduced Amp as successor product for agentic workflows
- Cody Enterprise remains fully supported and invested in
- Company strategy shifted from individual developers to enterprise-only
- Represents risk for organizations evaluating – product direction unclear
6. Windsurf (formerly Codeium)
Vendor: Windsurf / Cognition AI (acquired Dec 2025)
Pricing Tiers (March 2026):
| Tier | Price | Credits/mo | Key Differentiators |
|---|---|---|---|
| Free | $0 | 25 | Unlimited Tab completions, 1 App Deploy/day |
| Pro | $15/mo | 500 | All premium models, Previews, App Deploys |
| Teams | $30/user/mo | 500/user | Admin dashboard, analytics, priority support, pooled add-on credits |
| Enterprise | $60/user/mo | 500 (1,000 at 200+ seats) | RBAC, SSO/SCIM, longer contexts, hybrid/self-hosted deployment |
Enterprise Features:
- SSO/SCIM included in Enterprise; available on Teams for +$10/user/mo
- RBAC (Role-Based Access Control)
- Volume-based discounts at scale
- Hybrid and self-hosted deployment options (Enterprise only)
- Account management
- Add-on credits: $10/250 (Pro), $40/1,000 pooled (Teams/Enterprise)
Adoption Numbers:
- 1M+ developers (Apr 2025)
- ~$82M ARR at time of Cognition acquisition
- Grew from 10K users (early 2023) to 800K+ (early 2025)
Recent Changes (2025-2026):
- Acquired by Cognition AI for ~$250M (Dec 2025) after failed $3B OpenAI deal
- Rebranded from Codeium to Windsurf (Apr 2025)
- Introduced credit-based pricing model with SWE-1.5 fast agent model
- Cheapest full-featured Pro plan in the market at $15/mo
- Early adopter pricing locked through end of 2026 for existing subscribers
- Student pricing program available
- Ownership change creates uncertainty for long-term enterprise commitments
7. JetBrains AI Assistant
Vendor: JetBrains
Pricing Tiers (March 2026):
| Tier | Price | AI Credits/mo | Key Differentiators |
|---|---|---|---|
| AI Free | $0 | 3 credits/30 days | Basic AI features in JetBrains IDEs |
| AI Pro | $8/mo | 20 credits ($20 value) | Full AI Assistant + Junie agent |
| AI Ultimate | $30/mo | 70 credits ($35 value, $5 bonus) | Higher quotas, premium models |
| AI Enterprise | Custom | Custom | External LLM providers, on-prem models, centralized policies |
Enterprise Features:
- SSO provisioning and audit logging
- Connect approved external LLM providers (OpenAI, Anthropic, Google, Azure)
- On-premises model support (Hugging Face, llama.cpp, vLLM, LMStudio)
- JetBrains Mellum self-hosted engine
- Air-gapped environment support
- Profile-based policies, offline/local-only modes
- Part of JetBrains IDE Services platform
Adoption Numbers:
- No specific AI Assistant user numbers disclosed
- JetBrains IDEs have 16M+ users overall
- 93% of developers use AI tools regularly (JetBrains 2026 survey)
Recent Changes (2025-2026):
- Unified AI plans across AI Assistant and Junie (agentic coding)
- Introduced credit-based system (1 credit = $1 USD)
- Top-up credits available (valid 12 months)
- Added on-premises Mellum engine for self-hosted deployments
- AI Enterprise tier added with full isolation capabilities
8. Google Gemini Code Assist
Vendor: Google (Google Cloud)
Pricing Tiers (March 2026):
| Tier | Price | Key Limits | Key Differentiators |
|---|---|---|---|
| Individual (Free) | $0 | 6,000 code requests + 240 chats/day | No GCP project required, basic completions + chat |
| Standard | $22.80/user/mo | Higher quotas | Pooled usage, advanced security, via Google Cloud |
| Enterprise | $54/user/mo | Highest quotas | Code customization on private repos, Apigee + App Integration AI, GCP infra management |
Enterprise Features:
- IP indemnification for generated code (Google assumes responsibility)
- SOC 1/2/3, ISO 27001/27017/27018/27701 compliance
- VPC Service Controls, Private Google Access
- IAM with granular role-based access control
- Enterprise Access Controls
- Code customization based on private repositories
- No self-hosted option (Google Cloud only)
Adoption Numbers:
- No specific user numbers disclosed
- Integrated into Google Cloud ecosystem (available to all GCP customers)
- Google Developer Program premium unlocks higher quotas
Recent Changes (2025-2026):
- Free individual tier launched (no GCP project required)
- Powered by Gemini 2.5 model
- Added agentic chat and external service integration
- Standard and Enterprise tiers launched for organizations
- Gemini CLI introduced with separate quota system
- Integration with Google Developer Program for higher individual limits
Key Data Points
Market Overview
- AI coding tools market: $7.37B (2025), projected $8.5B (2026)
- 51% of professional developers use AI tools daily
- 93% of developers use AI tools regularly (JetBrains 2026 survey)
- GitHub Copilot holds ~42% market share
Pricing Trends
- Industry converging on credit/token-based pricing vs. flat-rate unlimited
- Enterprise tiers range from $19/user/mo (Copilot Business) to $60/user/mo (Windsurf Enterprise)
- Free tiers now standard but increasingly limited (Sourcegraph discontinued theirs entirely)
- Annual billing discounts of 15-20% are common
Enterprise Decision Factors
- IP indemnification: Explicitly offered by GitHub Copilot (Business/Enterprise), Google Gemini Code Assist (Standard/Enterprise), and Tabnine (license-safe AI)
- Self-hosted/air-gapped: Tabnine Enterprise and JetBrains AI Enterprise are the only options with true on-premises deployment
- Strongest governance: GitHub Copilot Enterprise (180-day audit logs, SCIM, SAML, policy controls)
- Best for AWS shops: Amazon Q Developer Pro ($19/user/mo) with deep AWS integration
- Best for GCP shops: Gemini Code Assist (free individual, $22.80-54/user/mo for orgs)
- Best value for small teams: Windsurf Pro at $15/mo or JetBrains AI Pro at $8/mo
- Highest risk: Sourcegraph Cody (product pivot) and Windsurf (acquisition uncertainty)
Adoption Rankings (estimated users, March 2026)
- GitHub Copilot: 20M+ total, 4.7M paid
- Cursor: 1M+ total, 360K+ paid
- Windsurf/Codeium: 1M+ total
- Tabnine: 1M+ active developers
- Amazon Q Developer: Not disclosed (thousands of enterprise teams)
- Google Gemini Code Assist: Not disclosed (bundled with GCP)
- JetBrains AI: Not disclosed (16M+ IDE users)
- Sourcegraph Cody: Not disclosed (enterprise only post-pivot)
Sources
-
Changes to Cody Free, Pro, and Enterprise Starter Plans (Sourcegraph Blog)
-
Gemini Code Assist Standard and Enterprise Overview (Google Cloud Docs)
-
Gemini Code Assist Security, Privacy & Compliance (Google Cloud Docs)
-
Tabnine Enterprise Context Engine Launch (GlobeNewswire, Feb 2026)
-
Cursor Enterprise Review: Pricing, Pros, Cons 2026 (Superblocks)
What This Means for Your Organization
The pricing data tells a clear story: the entry point for AI coding tools has converged around $19/seat/month (GitHub Copilot Business), but the total cost of a full AI stack can reach $168-199/seat/month when you layer in M365 Copilot, enterprise IDE features, and premium model access. Knowing where your organization sits on that spectrum – and where the diminishing returns begin – is a six-figure annual decision for any team of 100+ developers.
Three factors should drive your procurement: IP indemnification (only GitHub Copilot, Google Gemini Code Assist, and Tabnine explicitly offer it), air-gapped deployment (only Tabnine and JetBrains AI Enterprise support it), and governance tooling (GitHub Copilot Enterprise leads with 180-day audit logs and SCIM). If you are in a regulated industry – financial services, healthcare, legal, defense – these are not optional features. They are compliance requirements that narrow your shortlist before pricing even enters the conversation.
The market is also shifting beneath you. Credit-based pricing is replacing flat-rate unlimited plans at Cursor, Windsurf, and JetBrains. Sourcegraph discontinued its free and pro tiers entirely. Lock in enterprise agreements now while vendors are still competing aggressively for market share – the pricing power will shift toward vendors once adoption locks in.
Research conducted March 16, 2026. Last updated March 16, 2026. Pricing and features subject to change. Always verify with vendor pricing pages before making procurement decisions.
Created by Brandon Sneider | brandon@brandonsneider.com March 2026