- 20
- July
Claude Code is Claude AI's (Anthropic) agentic coding tool that runs from the command line and inside your IDE, with a cloud sandbox mode as well. It is not just a chat that autocompletes code — it is an assistant that takes action: it opens pull requests, runs tests, refactors across many files, and spins up parallel subagents. This article explains what Claude Code is, which plans include it and how much they cost (prices stay in USD throughout), the billing change that took effect in mid-2026, and how development teams actually use it.
In one line: Claude Code is an agentic coding tool included with every paid plan — Pro ($20/mo), Team Standard, Team Premium, Max and Enterprise. The difference between Team Standard ($20/seat/mo) and Team Premium ($100/seat/mo) is not whether you get Claude Code, but how much usage you get — Premium gives roughly 5× the usage of a Standard seat, for teams that use Claude Code heavily every day.
What Claude Code Is — and What It Can Do
Earlier coding assistants mostly "completed code line by line" inside a single editor pane. Claude Code is designed as an agent: it takes a plain-language task and carries out multiple steps in a real project. It reads the whole project structure, edits several files at once, runs terminal commands, and checks its own work against test results. The developer stays in control and approves every change, but the amount of work you can hand off in a single request grows substantially. The table below summarises what Claude Code can do.
| Capability | What it means |
|---|---|
| Edit across many files | Takes a task like "add feature X" and updates every relevant file, not just the one you have open. |
| Open a pull request | Creates a branch, writes commits, and opens a PR with a description so your team reviews it the normal way. |
| Run and fix tests | Runs the test suite, reads the failures, fixes the code, and re-runs until it passes. |
| Refactor a whole project | Renames, restructures, or moves logic across modules while preserving existing behaviour. |
| Parallel subagents | Splits a large task into several agents that work concurrently, then merges the results. |
| Cloud sandbox mode | Runs long tasks on the cloud in an isolated environment, in addition to running in your local terminal/IDE. |
Capabilities at this level explain why Claude Code sits in the group of agentic coding tools many teams are comparing (we cover the tooling angle in Codex vs Claude Code and using AI for code review).
Which Plans Include Claude Code — Pricing and What You Get
This is the heart of the article, because the biggest source of confusion is "which plan do I subscribe to in order to get Claude Code?" The key fact to remember: Claude Code is included with every paid plan — Pro, Team Standard, Team Premium, Max and Enterprise. What differs is not "included or not" but the usage limit: a Team Premium seat gives roughly 5× the usage of a Team Standard seat, plus Cowork. The table below compares the organisation and individual plans relevant to Claude Code (prices stay in USD as published by Anthropic).
| Plan | Price (USD) | Claude Code? | Usage / seat range |
|---|---|---|---|
| Team Standard | $20/seat/mo (annual) = $240/seat/yr | Included + Cowork | Standard usage (more than Pro) · min 2 seats, max 150 |
| Team Premium | $100/seat/mo (annual) = $1,200/seat/yr · monthly ~$125/seat (about 25% more than annual) | Included + Cowork | ~5× the usage of Standard · 2–150 seats · you can mix Standard + Premium in one org |
| Max (individual) | $100 and $200/mo (two tiers) | Included | For a single user, much higher usage than Pro |
| Pro (individual) | $20/mo | Included | For a single user, good for trials/light work |
| Enterprise | Custom pricing (contact sales) | Included | Min 20 seats, no maximum cap |
The clear takeaway for organisations: Team Premium, Team Standard and Enterprise all include Claude Code, so the decision is not "which plan has Claude Code" but "how heavily will our team use it." A team that uses Claude Code as a daily primary tool will hit a Standard seat's usage ceiling quickly and is better on Premium (~5× the usage), while occasional users are fine on Standard. Individuals who just want to try Claude Code can do so on the Pro plan (though Grand Linux Solution sells Team/Enterprise plans for organisations, not individual Pro/Max plans). For per-plan limits, see Claude plans and usage limits, and compare Max vs Team.
Two billing facts to know before you decide: (1) Annual = a one-year commitment — buying through Grand Linux Solution is annual, paid 100% upfront, with no monthly option and no mid-year refund, so estimate your seat count carefully first. (2) Since 15 June 2026 Anthropic split the billing — interactive Claude Code use in the terminal/IDE draws on your plan limits, while programmatic use (via the Agent SDK) draws from a separate, dollar-denominated API credit at full API rates. Teams planning automation should budget for that separately.
Interactive vs Programmatic — Where the Money Comes From
Because the 15 June 2026 billing split affects real budgets, the table below makes it clear that the two usage modes draw from different pockets.
| Usage type | Example | Charged to |
|---|---|---|
| Interactive | A developer typing commands to Claude Code in the terminal or IDE during day-to-day work | Your plan limits (Pro/Max/Team Premium/Enterprise) — no separate bill |
| Programmatic | Calling Claude Code via the Agent SDK to run in a CI/CD pipeline or an unattended automation script | A separate API credit in dollars at full API rates — must be topped up and budgeted separately |
A note for leaders setting budgets: if your team uses Claude Code as part of normal hands-on work, the cost stays within a predictable annual seat price. But once you build real automation (for example, an agent that runs on every code push), you need to set aside a separate API-credit budget, or the numbers can grow without you noticing.
Recommended Model for Coding: Claude Opus 4.8
As of July 2026, the recommended model for coding is Claude Opus 4.8 (released 28 May 2026). Anthropic reports Opus 4.8 at around 88.6% on SWE-bench Verified, a benchmark of fixing real software bugs.
Read this number carefully: the ~88.6% figure is vendor-reported (self-reported by Anthropic) and reaches us via secondary sources citing Anthropic — it is not an independent evaluation result. Treat it as "the developer's reported figure," not a neutral conclusion. The most reliable approach is to test against your own team's codebase and real tasks, because a vendor benchmark does not always equal performance on your actual workload.
How Development Teams Actually Use Claude Code
In practice, development teams don't use Claude Code to replace people — they use it to cut repetitive work and shorten well-defined tasks. Common examples include writing tests to cover legacy code that never had any, doing large refactors that people tend to postpone because they're risky and tedious, working through vulnerabilities a security scanner has flagged, explaining old undocumented code, and drafting a first PR for review. The important part: every change still passes through a developer's eyes and approval — the agent speeds up the drafting, but a human still decides.
For organisations getting started, a safe approach is to begin with a small number of Team Premium seats for a pilot team, measure against your own real work first, then expand. Because the Team seat structure lets you mix Standard and Premium in one organisation, the people who use Claude Code heavily every day take Premium seats (for ~5× the usage) while occasional users or general-chat users take Standard (which still includes Claude Code, just with a lower limit) — which helps control the budget (read more in What is Claude Team).
Buying Direct from Anthropic vs Through Grand Linux Solution
Thai organisations often ask, "If I can subscribe with Anthropic directly, why buy through an intermediary?" The straightforward answer: buying through Grand Linux Solution (GLS) is one option for organisations that want convenience around paperwork, tax, and Thai-language coordination. You decide entirely for yourself on price and value. The table below shows what each side handles.
| Item | Buy direct from Anthropic | Buy through Grand Linux Solution |
|---|---|---|
| Thai tax invoice (VAT 7%) | A foreign bill; no Thai tax invoice | Issues a complete Thai tax invoice |
| Withholding tax / Phor.Phor.36 | You handle imported-service tax yourself | Helps handle the paperwork and tax process under Thai rules |
| Quote / currency | In dollars, paid by an international card | Issues a Thai-language quote, paid in baht at a fixed exchange rate |
| Purchase order / renewal | You manage the PO process and renewal dates yourself | Handles PO issuance, billing, and renewal reminders |
| Coordination with Anthropic | You contact them in English yourself | Coordinates with Anthropic in English on your behalf |
What you get when buying through Grand Linux Solution: a Thai tax invoice (VAT 7%), handling of withholding tax and Phor.Phor.36, a Thai-language quote paid in baht, PO issuance plus billing and renewal management, and coordination with Anthropic in English on your behalf. Higher-cost services such as team training, onboarding, internal-system/SSO integration with your ERP, and deep technical support are paid optional add-ons separate from the licence fee — not bundled for free. We price them by actual scope so nothing is hidden inside the licence cost.
Data caveat: Claude is a cloud SaaS processed on Anthropic's infrastructure (US by default). For Team/Enterprise plans, data is not used for training by default, there is roughly 7-day API retention, and Zero Data Retention (ZDR) is available by addendum. Still, in principle the prompts and code you send leave your organisation to be processed in the cloud, so you should not put secrets, personal data (PDPA), or credentials into it without governance policies in place. Read more in Claude Team and PDPA and security and data governance.
Where Saeree ERP Fits In (Honestly)
To be candid, Claude Code is a tool for software development teams — it is not part of Saeree ERP, and the AI assistant inside the ERP itself is still in development (the training phase), not a usable feature yet. The honest connection is this: organisations with an in-house IT or development team maintaining back-office systems often use tools like Claude Code for scripting, reporting, or system integration work. In that role, GLS can supply Claude licences with Thai tax documentation alongside looking after the ERP system — from a single provider. Note that integrating Claude directly with an ERP is an integration engagement offered as a paid add-on, not an off-the-shelf feature.
"The more important question than 'how much does Claude Code cost' is 'how much real value will our team create with it' — because an annual licence is a one-year commitment. Start with a pilot team, measure against real work, then expand with confidence. There's no need to rush."
- Sureeraya Limpaibul, Grand Linux Solution
References
- Anthropic — Claude Code (product overview)
- Anthropic — Claude Code Documentation / Overview
- Anthropic — Pricing (Pro / Max / Team / Enterprise)
- Anthropic Support — Using Claude Code with Pro/Max plan
- Anthropic Support — Using Claude Code with Team/Enterprise plan
- Anthropic — Claude Opus 4.8 (vendor-reported SWE-bench figure)
Information last verified on 20 July 2026.
Interested in Claude licences for your development team?
Grand Linux Solution sells Claude Team and Enterprise licences with a Thai tax invoice, a Thai-language quote, payment in baht, and coordination with Anthropic on your behalf. Buying through us is one option for organisations that want convenience around paperwork and tax — you decide on price and value. Licences are annual, so consider it carefully first. There's no need to rush.
Request a QuoteTel 02-347-7730 | sale@grandlinux.com
