Claude Code
Released Feb 24, 2025
Anthropic's agentic coding tool: reads a codebase, edits files and runs commands, available in the terminal, IDE extensions, a desktop app and the browser.
- Interface
- CLI, IDE, DESKTOP, WEB
- Pricing
- Subscription
- Models
- Vendor's own models
- Best Terminal-Bench
- 64.85% (v4.0) · Claude Opus 5.5
- Open source
- No
- Licence
- Proprietary
Terminal-Benchv4.0 · per harness + model pairing
| Claude Opus 5.5 | 64.85% |
| Claude Sonnet 5.5 | 61.82% |
| Claude Fable 5.1 | 57.88% |
| Claude Opus 5 | 53.94% |
| Claude Fable 5 | 44.55% |
| GLM-5.3 | 41.82% |
| GLM-5.3-Flash | 35.76% |
| Qwen3.8-Max-0902 | 26.97% |
| Claude Opus 4.8 | 23.64% |
| Claude Sonnet 5 | 12.42% |
Strengths & weaknesses
Strengths
- #1 and #2 on Terminal-Bench 4.0 (Opus 5.5 64.8%, Sonnet 5.5 61.8%)
- Runs on terminal, VS Code, JetBrains, desktop app, web and mobile
- Subagents, skills, hooks and MCP integration
- Third-party providers supported on the CLI, VS Code and JetBrains
Weaknesses
- Not open source: the public repo is 'All rights reserved', under Anthropic's Commercial Terms
- Most surfaces require a paid Claude plan or Console account
Notable: appears on the official Terminal-Bench leaderboard (tbench.ai, Terminal-Bench 4.0 — the only version tbench.ai currently shows), best rank #1; only five distinct agents are listed there, so every one is top-10. Scores are TB 4.0 resolution rates (best per model across reasoning efforts), fetched 2026-10-14 — not comparable with older Terminal-Bench 2.x figures. releaseDate is the limited research preview announced with Claude 3.7 Sonnet. supportedModels='vendor' per Anthropic's docs (Claude models, incl. via Bedrock/Vertex-style third-party providers), although tbench.ai also lists Claude Code runs on non-Anthropic models (GLM-5.3, GLM-5.3-Flash, Qwen3.8-Max-0902) — included in scores since they are official leaderboard rows.