I am a heavy user of agentic coding tools, part of my job is to evaluate their capabilities and limitations. This is an experiment where I am tracking their capabilities, limitations and my overall impression of them.
Note: Some cells have more information when you hover over them.
Colour Key:
- Requires subscription
- Minor limitation
- Moderate
- Significant
- Major
| Claude Code | Cline | OpenCode | OpenAI Codex | Github Copilot | Google Antigravity | Cursor | Amazon Kiro | |
|---|---|---|---|---|---|---|---|---|
| Functionality | ||||||||
| Inference Providers & Model Selection | Any (works best with Anthropic models) | Any | Any | OpenAI models only, US based or BYO baseURL | US based with limited context window, can BYO | US based Gemini models only | US based models, limited by plan, BYO option but limits features | US based models |
| Sub-Agents | Yes | Yes | Yes | No | No | Partial / very limited | Yes (limited) | No |
| Custom Agents | Yes | Yes | Yes | No | Yes | Yes | No | No |
| Prompt templates | Yes | Yes | Yes | No | Yes | Partial | Yes | No |
| Coding abilities | Best | Very good | Good | Good | Lower end | Medium | Mid range | Mid range |
| Speed | Fastest | Fastest | Good | Good | Slow | Medium | Slow | Slow |
| MCP Spec alignment | High | Medium/High | Medium | Medium/High | High | Medium | Medium/Low | Low |
| Full model capability | Full | Full | Full | Full | Heavily Context Limited | Full | Context Limited | Context Limited |
| Operational management | ||||||||
| Consumption budget controls | With Enterprise | With Enterprise | No | Partial (basic project/key limits) | N/A | No | No (for BYO) | N/A |
| Developer experience | ||||||||
| User Interface | VSCode, IntelliJ IDE, TUI, Zed, CLI, CI/CD | VSCode IDE, IntelliJ, CLI | TUI, CLI | VSCode IDE, CLI | VSCode, Zed, CLI | Standalone IDE (Fork of VSCode), CLI | Standalone IDE (Fork of VSCode), CLI | Standalone IDE (Fork of VSCode), CLI |
| Source | Closed source | Open source | Open source | Open source | Open source | Closed source | Closed source | Closed source |
| Babysitting required | Very Low | Low | Medium/Low | Medium/Low | Very High | High | Very High | Very High |
| Plan/Act modes | Yes | Yes | Yes | No | Yes | Partial (doesn't work) | Yes | No |
| Workflow flexibility | High | High | High | High | Medium | Medium | Medium | Very Low |
| Community size | Very Large | Large | Medium | Medium | Large | Small | Large | Small |
| Configuration management | ||||||||
| File based configuration | Yes | No | Yes | Yes | Partial | Limited | No | Limited - no RBAC or rules |
| Per-project configuration (rules, MCP, RBAC) | Yes | In beta | Partial | Partial | Limited - no RBAC | Limited | No | Limited - no RBAC |
| Centralised configuration | Partial (manual - split across config files) | With enterprise / teams subscription | No | No | Partial | No | No (for BYO), Yes for US based | No |
| Automation | ||||||||
| Automation hooks | Yes | Yes | Yes (Plugins) | Yes | No | No | No | Yes |
| Headless / CLI / CI/CD mode | Yes | Yes | Yes | Yes | Yes | Yes | In beta | No |
| Automated Context Cleanup | Yes | Yes | Yes (Plugin) | No | Yes | No | No | |
| Access control | ||||||||
| RBAC - Commands | Yes | Partial (basic) | Yes | Partial | No | No (has basic allow list - easy to bypass) | No (has basic allow list - easy to bypass) | No (has basic allow list - easy to bypass) |
| RBAC - Web domains | Yes | With enterprise / teams subscription | Yes | No | No | No | No | No |
| RBAC - MCP tools | Yes | With enterprise / teams subscription | Yes | No | No | No | No | No |
| SSO | Yes | With enterprise / teams subscription | N/A | No | Yes | Yes | No (for BYO), Yes for US based | Yes |
| User Interface (Subjective) | ||||||||
| User Interface | Best | Best | Good | Good | Good | Good | Good | Average |
| Cost (Estimate) | ||||||||
| Monthly cost ($USD) estimate for agentic coding (used to the best of the tools limitations) | $100-$200 | $1200-$2200 | $1200-$2200 | $200+ | $100-$300 | $200+ | $200+ | $200+ |