Claude Code
Highest published score on a coding-agent test, from $20/month; OpenAI's Codex scores close for far less per task.
This product is made by Anthropic, whose Claude models help draft our reviews. How we keep that from tilting a score
Should you use it?
Pick it if
Developers who want a terminal agent on the strongest current coding model and already pay for Claude.
Best for: Developers who already pay for Claude and want the strongest current coding model (Coding agents Top 20)
The short version
- What this is
- Anthropic's coding agent: it reads your repository, edits files and runs commands from the terminal, IDE, desktop app or browser.
- Why it matters
- It has the highest published score on the one current independent coding-agent test, but costs more per task and shares limits with chat.
- Default model Opus 5.5 ranks first on arena.ai WebDev and Agent s3 s4 s8.
- Needs Claude Pro at $20/month; no free plan s5.
- Scores 70 for coding on our checklist: top test results (68 and 66), held back by cost, shared limits and the training setting s12 s5 s7.
Good at
- Runs Anthropic's Opus 5.5 by default on paid plans s3, the top model on arena.ai WebDev and Agent today s4 s8.
- Highest score of the rows Artificial Analysis shows on its Coding Agent Index v1.5: 68 with Sonnet 5.5 (max) s12.
- Maintained at a fast pace: three releases between 8 and 9 Oct 2026, latest v2.1.296 s10.
Weak at
- No free plan; the $20 Pro plan shares one usage pool with Claude chat, reset every five hours s5.
- Costly per task: $13.04 with Opus 5.5 (max), against $1.04 for Codex on GPT-6.1 Sol (xhigh), which scored 63 s12.
- Artificial Analysis's main index hides its values and doesn't list every agent, so the full field can't be seen s2.
Our take
The model underneath is the strongest on the public boards today: Opus 5.5 is the default on paid plans and ranks first on arena.ai WebDev s3 s4. On Artificial Analysis's Coding Agent Index v1.5, Claude Code scores 68 with Sonnet 5.5 (max) and 66 with Opus 5.5 (max), the highest of the rows the comparison pages show, against 63 for Codex on GPT-6.1 Sol and 54 for OpenCode on GLM-5.3 s12 s14. Version 1.5 has been in use since September 2026 s13, so these are current results. The main index hides its values s2, so we can't see every agent.
The score is 70, not higher, because of what it costs to use. There is no free plan, Pro at $20 shares one usage pool with Claude chat s5, and a task cost $13.04 with Opus 5.5 (max) against $1.04 for Codex at a score of 63 s12. On personal plans, sessions train Anthropic's models if the setting is on s7. If you want a non-Anthropic option at the same price, Codex CLI with ChatGPT Plus is the closest scorer s12. This product is made by Anthropic, whose Claude models help draft our reviews.
Score
A score from 0 to 100 per job, against everything available today. Overall 70 across 1 job, the average of the job scores.
New to this? What the scores and labels mean
- 0 to 100
- Our score for one job, like writing or coding, compared with everything else out today. 100 is the best available today. For AI models we build it from public data, such as Arena's blind votes and Epoch AI's independent tests; apps and tools are scored on our own checklist. The same tool can score differently for different jobs, and where two likely ranges overlap, we call it a tie.
- Our verdict
- Use it: a top choice for most people. Worth trying: good, and worth a go. Depends: good for some people or one kind of job. Wait: too early to tell, or not out for everyone yet. Skip: something else does the job better.
- First look
- Just released. We say what it is and whether to bother, and give it a score once public results exist.
- No score yet
- None of the public sources we use covers it yet, or the only evidence is from a source whose numbers we can't use, so we describe it in words instead.
- Best buy
- Nothing cheaper scores as well for that job.
- Best free
- The best score for that job that costs nothing, through a free plan or by running on your own computer.
- Pay as you go
- No monthly plan: you pay for each use, usually through a developer account.
- Checked
- The day we last confirmed the score and the price against the sources. Prices change often, so check the date.
- AI model
- The engine underneath, like GPT or Claude. You use it through an app or, if you code, directly.
- Open source
- Free to download and run on your own computer, so nothing leaves your machine.
Our checklist score, going by an independent test, a public leaderboard and the maker’s pages.
Our checklist
2 sources
Sources and notes
Independent testPublic leaderboardMaker’s pagesChecklist 7/10: does the job 2 (Artificial Analysis Coding Agent Index v1.5, in use since September 2026 and the first version with Terminal-Bench 4.0 s13, scores it 68 with Sonnet 5.5 (max) and 66 with Opus 5.5 (max), the highest of the rows it shows, against 63 for Codex and 54 for OpenCode s12 s14); underlying quality 2 (default Opus 5.5 is arena.ai WebDev rank 1 of 142, claude-opus-5.5-max s3 s4); value 1 (Pro $20/month, twice Copilot Pro s5 s6; $13.04 a task with Opus 5.5 (max) s12); privacy 1 (personal plans train when the model-improvement setting is on; the docs do not state the default s7); limits 1 (Pro shares one usage pool with chat, reset every five hours s5).
Alternatives
Better or cheaper for coding, from our scores, then the ones we'd also look at. The same kind of product comes first.
- OpenCodeBetterCoding agent80FreeScores 80 in coding, against 70 here; free. Side by side
- OpenAI Codex CLICoding agent70$20 a monthOpen-source terminal agent included with ChatGPT Plus at the same $20. Side by side
- GitHub Copilot (including Copilot CLI and cloud agent)Coding agentFirst look: Worth trying$10 a monthCheapest paid option at $10/month, with a CLI and a cloud agent that opens pull requests. Side by side
- CursorCoding agentFirst look: Worth trying$8 a monthAn editor rather than a terminal tool, with a cheap in-house model. Side by side
Value for money
Where it sits in coding: higher is a better score, further left is cheaper. Green dots are the Best buys.
Evidence
Everything behind the call and the scores, folded so the page stays short. Spotted a wrong price, score or link? Report a mistake.
How we scored this
The coding score comes from our own checklist of the product.
Independent testPublic leaderboardMaker’s pages
This product is made by Anthropic, whose Claude models help draft our reviews.
Reviews are drafted with help from AI models, including Anthropic's Claude. Model scores come only from outside leaderboards and independent evaluations, by a written rule.
No one paid for this review or gave us free access, and we earn nothing if you sign up. How we score
Sources 14
- s1Artificial Analysis: Cursor Composer 2.5 on the Coding Agent Index (20 May 2026)
- s2Artificial Analysis: Coding agents leaderboard
- s3Claude Code docs: Model configuration
- s4arena.ai Code Arena: WebDev leaderboard
- s5Claude pricing
- s6GitHub Copilot plans
- s7Claude Code docs: Data usage
- s8arena.ai Agent leaderboard
- s9Claude Code docs: Overview
- s10Claude Code releases on GitHub
- s11Anthropic: Introducing Claude 4 (Claude Code generally available)
- s12Artificial Analysis: Claude Code vs Codex, Coding Agent Index v1.5
- s13Artificial Analysis: Coding Agent Index v1.5 methodology
- s14Artificial Analysis: Claude Code vs OpenCode, Coding Agent Index v1.5
Something here out of date or wrong? Report a mistake and we will check it against the source.
History 6 changes
Every change to a rating, the price or our call, with the reason. Scores are against what exists on the day, so they drift down as better things ship. Older entries come from before we moved to scores.
- Plans: Max 20x is now called Ultra ($200/month); Max ($100/month) is five times Pro usage every 5 hours and every week. Prices unchanged; no score changes.
- First look. Independent agent results are older than 90 days or test older models, so not rated yet.
- Added Artificial Analysis Coding Agent Index v1.5 scores: 68 with Sonnet 5.5 (max), 66 with Opus 5.5 (max). No per-result date and the main index hides values, so still a First look.
- New: First look to ScoredNew: Coding ratedRated for coding by the checklist. Artificial Analysis's methodology page shows v1.5 has been in use since September 2026 and is the first version with Terminal-Bench 4.0, so its scores fall inside the 90-day window. Agents not rated yet.
- Audit: verdict line and why now say highest published score, not top scorer, since Artificial Analysis's main index hides its values and the full field can't be seen. Scores re-confirmed. Rating unchanged.
- Verdict line now says OpenAI's Codex, since Artificial Analysis does not name which Codex surface it tested.
Use it well 4 guides on this site
Pages on this site that help with the jobs above.