The best Claude Code alternative for most developers is OpenAI Codex, included in ChatGPT Plus at $20 a month. OpenCode is the best free, open-source option, GitHub Copilot CLI is the cheapest paid plan with Claude models at $10, and Grok Build comes with an $8 X Subscription.
Last verified: November 6, 2026
The 8 Best Claude Code Alternatives
- OpenAI Codex: closest match to Claude Code
- OpenCode: best open-source agent for any model
- GitHub Copilot CLI: cheapest paid plan with Claude models
- Grok Build: best $8 terminal agent bundle
- Aider: best open-source agent built around git
- Cline: best open-source agent inside VS Code
- Antigravity CLI: Google’s terminal agent, which replaced Gemini CLI
- Cursor: best if you’d rather work in an AI editor
These Claude Code competitors fall into three groups: subscription agents (Codex, Copilot CLI, Grok Build and Antigravity CLI), open-source tools like Claude Code that you connect to any model (OpenCode, Aider and Cline), and AI editors (Cursor).
How we compared them: This is a research comparison based on pricing pages, documentation and help articles, not a hands-on test, so our picks are “best fit” calls rather than benchmark results. We ranked each tool on four things, in this order: how closely it matches Claude Code’s agent workflow, its monthly cost for regular use, the choice of models, and whether there’s a free or open-source option. On November 6, 2026, we checked Claude’s pricing and help pages, Anthropic’s Claude Code cost and data docs, OpenAI’s Codex documentation, GitHub’s Copilot plans and Copilot CLI docs, Cursor’s pricing, xAI’s Grok Build docs, OpenCode’s docs, Cline’s docs, Z.ai’s GLM Coding Plan docs and the Gemini CLI change notice. Plan details for Grok Build, Cline, Antigravity CLI and the “Also Considered” tools come from dated independent sources, which we name. All prices are in US dollars. SmartWorkflowLab has no sponsorship or affiliate deal with any tool in this list.
Claude Code Pricing in 2026: The Baseline
What is Claude Code? Claude Code is Anthropic’s coding agent, which works in your terminal, editor or the Claude apps, reads your codebase, edits files and runs commands to complete tasks you describe.
According to Claude’s pricing page and Anthropic’s Ultra plan article:
| Plan | Price | Usage | Includes Claude Code |
|---|---|---|---|
| Free | $0 | Limited | ✗ |
| Pro | $20 a month, or $17 billed yearly | 1x | ✓ |
| Max | $100 a month | 5x Pro | ✓ |
| Ultra | $200 a month | 10x Pro every week, no five-hour limit for verified accounts | ✓ |
| Team, standard seat | $20 a month yearly, $25 monthly | More than Pro | ✓ |
| Team, premium seat | $100 a month yearly, $125 monthly | 5x a standard seat | ✓ |
Anthropic renamed its Max 20x plan Ultra and moved existing subscribers over at the same price. Max and Ultra now also include $100 and $200 a month of API credits, but those credits can’t be used in interactive Claude Code, according to Anthropic’s API credits article.
Three things push developers toward Claude Code alternatives:
- Shared limits. Pro and Max have a five-hour session limit and a weekly limit, and “these limits are shared across Claude and Claude Code,” according to Anthropic’s Claude Code plan article. A long chat can eat into your coding time.
- A big price jump. The step after Pro at $20 is Max at $100.
- Claude models, officially. Claude Code is built for Anthropic’s models. Running it with other companies’ models takes a third-party endpoint that Anthropic doesn’t support, covered in the next section.
Verdict: Claude Pro is fair value at $20 if Claude’s models suit you and you rarely hit limits. If you do, or you want other models, read the next section before you switch tools.
Can You Keep Claude Code and Change the Plan or Model?
Yes. Three options keep Claude Code itself.
1. Pay per token through the Claude Console. If you sign in with a Claude Console account instead of a subscription, Claude Code is billed per API token, with workspace spend limits you set instead of the subscription’s usage windows, according to Anthropic’s cost documentation. It isn’t cheap for daily use. Across enterprise deployments, Anthropic puts the average at about $13 per developer per active day, or $150 to $250 a month, with 90% of users under $30 a day. API billing suits light or occasional use, while a subscription is usually better value if you code with it every day.
2. Turn on usage credits. Pro and Max subscribers can turn on usage credits to keep working past their plan’s limit, with an optional monthly spend limit, according to the same documentation.
3. Point Claude Code at another model. Several model providers offer Anthropic-compatible endpoints that Claude Code can use when you set environment variables such as ANTHROPIC_BASE_URL and ANTHROPIC_AUTH_TOKEN. A community list on GitHub includes Z.ai (GLM models), Moonshot (Kimi), DeepSeek, MiniMax, Alibaba (Qwen), OpenRouter and local models through Ollama. Z.ai sells a GLM Coding Plan from $18 a month that lists Claude Code as a supported tool, with 2,000 credits every five hours on its Lite plan, according to Z.ai’s documentation.
Keep three caveats in mind with option 3. Anthropic doesn’t support these setups, some features may behave differently with other models, and your code goes to that provider under its own terms.
Companies can also run Claude Code through Amazon Bedrock, Google Cloud’s Agent Platform or Microsoft Foundry, billed per token to their cloud account.
Verdict: If you like Claude Code’s workflow and only mind the limits, try usage credits, API billing or a GLM Coding Plan before you move to a new tool.
Claude Code Alternatives at a Glance
| Tool | Type | Models | Open source | Entry price |
|---|---|---|---|---|
| OpenAI Codex | Terminal, editor, cloud | OpenAI’s GPT models | ✓ (CLI) | ChatGPT Plus: $20 a month |
| OpenCode | Terminal, editor, desktop | 75+ providers, including local | ✓ | Free, plus model costs |
| GitHub Copilot CLI | Terminal, editor, cloud | Several providers, including Claude | ✗ | Pro: $10 a month |
| Grok Build | Terminal | Grok models | ✓ (client) | X Subscription Plus: $8 a month |
| Aider | Terminal | Almost any model, including local | ✓ | Free, plus model costs |
| Cline | VS Code, CLI, desktop | Your own API keys or Cline credits | ✓ | Free, plus model costs |
| Antigravity CLI | Terminal | Gemini models by default | ✗ | Free tier with a weekly cap |
| Cursor | AI code editor | Grok and Composer; others on Premium | ✗ | Plus: $8 a month |
What Each Option Costs
| Monthly budget | What you can get |
|---|---|
| $0 | OpenCode, Aider or Cline with a local model; Copilot Free; Codex quick tasks in ChatGPT Free; Antigravity’s free tier |
| $8 | Grok Build with X Subscription Plus; Cursor Plus; Codex light tasks in ChatGPT Go |
| $10 | GitHub Copilot Pro, with 1,500 AI credits a month; ClinePass at $9.99 for open-weight models in Cline |
| $18 | Z.ai’s GLM Coding Plan Lite, usable inside Claude Code |
| $20 | Codex in ChatGPT Plus; Claude Code in Claude Pro |
| $30 | Cursor Premium, plus API charges for Claude or GPT models |
| $39 | GitHub Copilot Pro+, with 7,000 AI credits a month |
| $100 | Claude Max; ChatGPT Pro at its $100 tier; Copilot Max with 20,000 AI credits |
The open-source tools cost nothing themselves, but hosted models are billed per token. For scale, Anthropic’s own average for Claude Code on API billing is about $13 per developer per active day, so heavy daily use with per-token billing can cost far more than a $20 subscription.
What “$0 with a local model” really needs. Local coding models need a capable computer. OpenAI’s gpt-oss 20B needs about 16GB of memory, while Qwen3-Coder 30B and Devstral 24B want a 24GB to 32GB graphics card or a Mac with 32GB, according to Morph’s Ollama model guide (updated October 16, 2026). The same guide lists Devstral 24B at 46.8% on SWE-Bench Verified, a common coding benchmark, and notes that Ollama’s default 4,096-token context must be raised for agent work.
Features Compared: MCP, Instruction Files and Cloud Tasks
MCP (Model Context Protocol) servers connect an agent to outside tools such as databases and issue trackers. AGENTS.md is an open instruction-file format that many agents read the way Claude Code reads CLAUDE.md.
| Tool | MCP servers | Instruction files it reads | Windows | Cloud tasks |
|---|---|---|---|---|
| Claude Code (for reference) | ✓ | CLAUDE.md; AGENTS.md through an import | ✓ | ✓ |
| OpenAI Codex | ✓ | AGENTS.md | ✓ | ✓ |
| OpenCode | ✓ | AGENTS.md, or CLAUDE.md as a fallback | ✓ (WSL recommended) | ✗, runs on your machine |
| GitHub Copilot CLI | ✓ | AGENTS.md, CLAUDE.md, GEMINI.md | ✓ | ✓ (paid plans) |
| Grok Build | ✓ | AGENTS.md | ✓ | Headless mode for scripts |
| Aider | ✗ | AGENTS.md through a config setting | ✓ | ✗, runs on your machine |
| Cline | ✓ | AGENTS.md and .clinerules | ✓ | ✗, runs on your machine |
| Antigravity CLI | ✓ | AGENTS.md and GEMINI.md | ✓ | ✗, runs on your machine |
| Cursor | ✓ | AGENTS.md | ✓ | ✓ (Premium) |
Sources: each tool’s own documentation, plus an independent MCP support tracker and Unblocked’s AGENTS.md guide (August 14, 2026) for Codex, Cursor, Aider and Claude Code.
How to Move Your Claude Code Setup
- Instructions. Copy your CLAUDE.md to AGENTS.md for Codex, Cursor, Cline, Grok Build and Antigravity CLI. OpenCode reads CLAUDE.md automatically when no AGENTS.md exists, along with
~/.claude/CLAUDE.mdand~/.claude/skills, according to OpenCode’s rules docs. Copilot CLI reads CLAUDE.md and.claude/CLAUDE.mddirectly, according to GitHub’s custom instructions docs. Aider needs areadentry for AGENTS.md in.aider.conf.yml. - One file for both. To keep using Claude Code alongside another tool, make AGENTS.md your main file and start CLAUDE.md with the line
@AGENTS.md, a workaround from Anthropic’s docs quoted in Unblocked’s guide. - MCP servers. Add each server again in the new tool’s own config file. Antigravity CLI, for example, moves MCP settings into a separate
mcp_config.json. - Skills, hooks and subagents. Copilot CLI supports skills, hooks and custom agents that run as subagents, according to GitHub’s customization overview. Custom slash commands usually need rewriting in each tool’s format.
1. OpenAI Codex: Closest Match to Claude Code
What is OpenAI Codex? Codex is OpenAI’s coding agent, which works in a command-line tool, an editor extension, a desktop app, the web and on iOS, and comes with every ChatGPT plan.
Best for: developers who already pay for ChatGPT Plus or want OpenAI’s models in a Claude Code-style agent.
Codex works the same way as Claude Code: you describe a task, it reads your project, edits files, runs commands and shows you the result. It can also run tasks in the cloud while you do something else. Its command-line tool is open source on GitHub under the Apache 2.0 license and installs on Windows as well as macOS and Linux.
According to OpenAI’s Codex pricing documentation:
| Plan | Price | Codex use |
|---|---|---|
| Free | $0 | Quick coding tasks |
| Go | $8 a month | Lightweight tasks with GPT-6 Luna |
| Plus | $20 a month | An estimated 15 to 160 GPT-6.1 Sol messages every five hours |
| Pro | $100, $200 or $500 a month | Maximum Codex tasks, no five-hour limit |
OpenAI calls the Plus ranges estimates, says weekly limits may also apply, and sells credits for extra use. Plus also includes automatic code review and a Slack integration. For a closer look at how Codex compares with an AI editor, see our Cursor Pro vs ChatGPT Plus comparison.
Downsides vs Claude Code
- OpenAI models only, so no Claude.
- Five-hour limits on Plus, like Claude Pro.
- Usage ranges are wide estimates, not fixed numbers.
Verdict: Start with Codex if you want the most Claude Code-like experience under a different subscription.
2. OpenCode: Best Open-Source Agent for Any Model
What is OpenCode? OpenCode is a free, open-source AI coding agent that runs in the terminal, in editors or as a desktop app, and connects to almost any model provider.
Best for: developers who want Claude Code’s workflow without being tied to one company’s models or limits.
OpenCode is the most direct open-source replacement for Claude Code. It supports “75+ LLM providers,” including local models through Ollama and LM Studio, and you can sign in with a GitHub Copilot subscription, according to OpenCode’s providers docs.
One catch for Claude users: OpenCode doesn’t accept Claude Pro or Max logins, and its docs say Anthropic prohibits this. You use Claude models through an Anthropic API key instead.
Pricing: OpenCode itself is free. You can pay your model provider directly or buy pay-as-you-go credits from OpenCode’s own model service. That service sells at cost and adds a processing fee of 4.6% plus $0.31 to each credit purchase, according to OpenCode’s docs.
Downsides vs Claude Code
- Per-token billing for most hosted models, which can cost more than $20 a month with heavy use.
- More setup: you choose and configure models yourself.
- No Claude subscription login.
Verdict: OpenCode is the pick for model freedom, and it reads your existing CLAUDE.md, so switching takes minutes.
3. GitHub Copilot CLI: Cheapest Paid Plan With Claude Models
What is GitHub Copilot CLI? Copilot CLI brings GitHub Copilot’s agent to the terminal, alongside Copilot in your editor and its coding agent that works on GitHub issues in the cloud.
Best for: budget-minded developers who want Claude and other models in one plan, especially if they already use GitHub.
“All plans include Copilot CLI and the Copilot app,” according to GitHub’s plans documentation. Copilot CLI supports MCP servers, skills, hooks and custom agents that run as subagents, and it reads AGENTS.md, CLAUDE.md and GEMINI.md files.
| Plan | Price | AI credits a month | Cloud coding agent |
|---|---|---|---|
| Free | $0 | A small allowance | ✗ |
| Pro | $10 a month | 1,500 | ✓ |
| Pro+ | $39 a month | 7,000 | ✓ |
| Max | $100 a month | 20,000 | ✓ |
| Business | $19 per seat | 1,900 per user, pooled | ✓ |
| Enterprise | $39 per seat | 3,900 per user, pooled | ✓ |
Business and Enterprise usage beyond the pool costs $0.01 per AI credit.
Downsides vs Claude Code
- You manage a monthly credit balance instead of time-based limits.
- The cloud coding agent needs a paid plan.
- On Free, Pro and Pro+, GitHub may use your interaction data for AI training from April 24, 2026 unless you opt out, according to Help Net Security.
Verdict: For $10, Copilot Pro is the cheapest way to keep Claude models in a paid agent plan.
4. Grok Build: Best $8 Terminal Agent Bundle
What is Grok Build? Grok Build is a terminal coding agent from SpaceXSI, the company behind Grok, which runs interactively, headlessly in scripts or inside editors through the Agent Client Protocol.
Best for: developers who want a paid terminal agent, an AI editor and Grok chat for $8 a month.
Grok Build supports MCP servers, AGENTS.md and subagents, and installs on macOS, Linux and Windows, according to xAI’s Grok Build docs. Its client has been open source under the Apache 2.0 license since July 2026, while the Grok models stay closed, according to AIreiter (updated July 29, 2026).
Pricing: Grok Build is free to try with limited usage. For more, new customers buy The X Subscription: Plus at $8 a month, Premium at $30, Super at $100 and Ultra at $200. Each plan’s Grok usage pool is shared with Cursor, according to CodeAgentSwarm’s pricing guide (updated October 31, 2026).
At $8, it ties with ChatGPT Go and Cursor Plus. The difference is the bundle: the same $8 also covers Cursor Plus, Grok chat and X Premium perks, while ChatGPT Go limits Codex to lightweight tasks.
Downsides vs Claude Code
- Grok models only through the subscription, with no Claude or GPT.
- A newer tool with less independent coverage.
- The usage pool is shared with Grok chat and Cursor.
Verdict: Grok Build is the best-value bundle if you can work with Grok’s models.
5. Aider: Best Open-Source Agent Built Around Git
What is Aider? Aider is a free, open-source AI pair programmer that runs in your terminal and works with almost any model.
Best for: developers who want every AI change saved as a reviewable git commit.
Aider “automatically commits changes with sensible commit messages,” so you can review, diff and undo every AI change with normal git tools, according to Aider’s site. It can connect to almost any model, including Claude through an API key and local models.
Pricing: Aider is free; you pay your model provider or run a local model.
Downsides vs Claude Code
- Less autonomous: it’s designed as a pair programmer you steer, not an agent you leave running.
- No built-in MCP support, according to the MCP support tracker.
- Setup through API keys and config files.
Verdict: Aider suits careful developers who value control and a clean git history over autonomy.
6. Cline: Best Open-Source Agent Inside VS Code
What is Cline? Cline is a free, open-source AI coding agent available as a VS Code extension, a command-line tool and a desktop app for macOS, Windows and Linux.
Best for: developers who want a Claude Code-style agent inside their editor and like approving each step.
Cline plans tasks, edits files and runs commands, and it supports MCP servers, according to Cline’s MCP docs. It also reads AGENTS.md alongside its own .clinerules files.
Pricing: Cline is free. You pay with your own API keys or with Cline credits sold at cost, or choose ClinePass at $9.99 a month for a curated set of open-weight models with five-hour, weekly and monthly quotas. The JetBrains extension is part of the Enterprise plan, according to UsagePricing’s Cline profile (captured November 2, 2026).
Downsides vs Claude Code
- Pay-per-token costs with your own keys.
- JetBrains support only on Enterprise.
- Reviewing each step is safer but slower.
Verdict: Choose Cline if you’d rather see and approve each step inside your editor.
7. Antigravity CLI: Google’s Terminal Agent
What is Antigravity CLI? Antigravity CLI, run with the agy command, is Google’s agent-first terminal tool, which shares its backend with the Antigravity desktop app.
Best for: Google AI subscribers and fans of Gemini models.
Google moved Google AI Pro, Google AI Ultra and free individual accounts from Gemini CLI to Antigravity CLI on June 18, 2026, according to the Gemini CLI maintainers. Its installer imports your Gemini CLI skills, MCP servers and agent profiles. Antigravity CLI reads AGENTS.md and GEMINI.md, supports MCP servers and runs on Windows, but unlike Gemini CLI it’s closed source, according to Avinash Sangle’s migration guide (May 21, 2026).
Pricing: The same guide lists a free tier with a tight weekly cap, and Google AI Pro at $19.99 a month with 1,000 credits a month. Check your current quota, as these figures date from launch.
Downsides vs Claude Code
- Closed source, unlike Gemini CLI.
- A tight weekly cap on the free tier.
- Limits now use five-hour and weekly compute windows instead of Gemini CLI’s daily request count, according to Machine Herald (June 1, 2026).
Verdict: Worth a try if you already pay for Google AI Pro, but check your quota before you rely on it.
8. Cursor: Best if You’d Rather Work in an AI Editor
What is Cursor? Cursor is an AI code editor built on VS Code, with Tab autocomplete, an in-editor agent and Cloud Agents on higher plans.
Best for: developers who used Claude Code mostly for edits they’d rather see inline as they type.
Cursor’s plans changed with The X Subscription. Plus at $8 a month runs Grok and Composer models. Premium at $30 adds Auto mode, Cloud Agents and third-party models such as Claude, billed at API list price from prepaid credit, according to Cursor’s pricing page.
Downsides vs Claude Code
- Claude models cost extra, on top of Premium’s $30.
- It’s an editor, not a terminal agent.
- Usage comes from a weekly pool shared with Grok.
Verdict: Cursor wins if you want AI in your editor as you type. For other editors, see our Cursor alternatives guide.
Also Considered
- Amp is Sourcegraph’s closed-source coding agent, with models from OpenAI, Anthropic and Google, a free Hobby plan with your own keys and an Individual plan from $20 a month, as of September 25, 2026.
- Kilo Code is an open-source agent for VS Code, JetBrains and the terminal. It’s free with your own keys, sells model credits at provider rates and offers Kilo Pass at $19 a month. Anaconda acquired Kilo in July 2026.
- Kiro is a spec-first coding agent billed through AWS, with a free plan of 50 credits a month and Pro at $20 for 1,000 credits, as of September 2026.
- Devin Desktop, the editor formerly called Windsurf, is included in Devin’s Pro plan at $20 a month, as of October 30, 2026.
- Goose is an open-source AI agent from Block, with a desktop app and a CLI, that works with more than 15 model providers.
- Zed is a fast editor that’s free with your own API keys, or $10 a month for Pro with $5 of hosted model credit.
Roo Code, once a popular Cline fork, shut down on May 15, 2026, according to Morph, so we left it out.
Privacy: Where Your Code Goes
- Claude Code on Pro or Max: Anthropic trains on your sessions only if you turn that setting on. Data is kept for five years with the setting on and 30 days with it off. Team, Enterprise and API use isn’t used for training, according to Anthropic’s data usage docs.
- Codex: ChatGPT’s Data Controls settings let you turn off model training, according to OpenAI’s help center.
- GitHub Copilot: Free, Pro and Pro+ interaction data may be used for training unless you opt out. Business and Enterprise data isn’t.
- Third-party endpoints and open-source tools: your code goes to whichever model provider you connect, under its terms.
- Local models: your code stays on your computer.
Before you send client code to any AI tool, read our guide on whether it’s safe to put client information into ChatGPT.
Which Tool Like Claude Code Should You Choose?
- “I like Claude Code but keep hitting limits.” Turn on usage credits, switch to API billing, or try a GLM Coding Plan inside Claude Code.
- “I want a second subscription for overflow work.” Add Codex with ChatGPT Plus for $20.
- “I want to keep using Claude models in another tool.” Copilot Pro at $10, or OpenCode, Aider or Cline with an Anthropic API key.
- “I want to spend less.” Copilot Pro at $10 or Grok Build at $8.
- “I want free and open source.” OpenCode or Aider with a local model, if your computer can run one.
- “I’d rather work in an editor.” Cursor, Cline or Devin Desktop.
For more on Anthropic’s plans, see our Claude directory page.
Frequently Asked Questions
What is the best alternative to Claude Code?
OpenAI Codex is the closest alternative for most developers, because it’s a terminal and cloud coding agent included in ChatGPT Plus at $20 a month, the same price as Claude Pro. OpenCode is the best open-source alternative, because it supports more than 75 model providers, including local models.
Is there a free alternative to Claude Code?
Yes. OpenCode, Aider and Cline are free, open-source tools where you pay only for the AI model you connect, or nothing if you run a local model on a capable computer. GitHub Copilot Free includes Copilot CLI with a small monthly allowance, and ChatGPT Free includes Codex for quick tasks.
How much does Claude Code cost in 2026?
Claude Code is included with Claude Pro at $20 a month, or $17 a month billed yearly, Max at $100 a month with 5x Pro’s usage and Ultra at $200 a month with 10x Pro’s weekly usage. You can also pay per token through the Claude Console. The Free plan doesn’t include Claude Code.
Why do people look for Claude Code alternatives?
The most common reasons are usage limits, price and model choice. Claude Pro has a five-hour limit and a weekly limit shared between Claude chat and Claude Code, the next step up costs $100 a month, and Claude Code is built for Anthropic’s Claude models.
Can I use Claude models without Claude Code?
Yes. OpenCode, Aider and Cline connect to Claude through an Anthropic API key, billed per token, though OpenCode doesn’t accept Claude Pro or Max logins. GitHub Copilot includes Claude models within its AI credits, and Cursor offers them on Premium at $30 a month, billed at API prices.
Can I use Claude Code with other models?
Not officially. Providers including Z.ai, Moonshot and DeepSeek offer Anthropic-compatible endpoints that Claude Code can use when you change its base URL setting, and Z.ai’s GLM Coding Plan, from $18 a month, lists Claude Code as a supported tool. Anthropic doesn’t support these setups, and your code goes to that provider.
Is OpenAI Codex better than Claude Code?
Neither is better for everyone. Codex comes with ChatGPT Plus at $20 and adds ChatGPT’s research and image tools, while Claude Code comes with Claude Pro at the same price and uses Anthropic’s models. Both have five-hour limits, so the fairest test is to run the same task in each on your own code.
Sources
- Anthropic: Claude pricing, Ultra plan article, Claude Code plan article, API credits article, Claude Code cost docs and data usage docs, checked November 6, 2026
- OpenAI: Codex pricing documentation, Codex on GitHub and data controls help article, checked November 6, 2026
- GitHub: Copilot plans, Copilot CLI customization overview and custom instructions, checked November 6, 2026; Help Net Security, March 26, 2026 (independent source)
- Grok Build and Cursor: xAI’s Grok Build docs and Cursor pricing, checked November 6, 2026; CodeAgentSwarm, updated October 31, 2026, and AIreiter, updated July 29, 2026 (independent sources)
- OpenCode: providers docs, rules docs, Windows docs and pricing docs, checked November 6, 2026
- Aider: Aider’s site, checked November 6, 2026
- Cline: Cline’s MCP docs, checked November 6, 2026; UsagePricing, captured November 2, 2026 (independent source)
- Google: Gemini CLI change notice, June 18, 2026; Avinash Sangle’s migration guide, May 21, 2026, and Machine Herald, June 1, 2026 (independent sources)
- Other models in Claude Code: Z.ai’s GLM Coding Plan docs, checked November 6, 2026; cc-compatible-models on GitHub (community list)
- Features and local models: MCP support tracker (independent source); Unblocked’s AGENTS.md guide, August 14, 2026, and Morph’s Ollama model guide, updated October 16, 2026 (independent sources)
- Also Considered: Amplifying for Amp and Kilo Code, checked September 25, 2026; Morph on Kiro, September 30, 2026; GPTPrompts on Devin, October 30, 2026; Morph on Roo Code, October 25, 2026 (independent sources); Goose on GitHub and Zed pricing, checked November 6, 2026
