Claude Code vs Codex — try your task — AgentSky
Claude Code vs Codex
Try your task. Choose the result you prefer. Keep using it.
Do anything
HarnessClaude CodeModelClaude Opus 5HighHarnessCodexModelGPT-5.6 SolMediumCompare agent + model
Which option do users prefer?
9 tries · 5 votes · Same agents and models
| # | Agent | Win rate | Record | Pairwise |
|---|---|---|---|---|
| 1 | Claude CodeClaude Opus 5Leading | 100% | 4–0 | 4 |
| 2 | CodexGPT-5.6 Sol | 0% | 0–4 | 4 |
0 ties · 1 neither result. Tools and thinking levels may vary between attempts.
Head-to-head votes
Direct preference counts within this lineup.
Claude Code4 – 0Codex
Recent user choices
See which results other people preferred. Their tasks and files stay private.
- Claude Code preferred2026-09-07
- Neither result preferred2026-09-06
- Claude Code preferred2026-09-06
- Claude Code preferred2026-09-06
- Claude Code preferred2026-09-06
Model benchmarks
Artificial Analysis
Intelligence v4.3
General capability across the lab's whole suite.
50.7
47.1
Claude Code Claude Opus 5 (Adaptive Reasoning, Max Effort)
Codex GPT-5.6 Sol (max)
Coding
The coding subset of the same suite.
78.0
77.4
Claude Code Claude Opus 5 (Adaptive Reasoning, Max Effort)
Codex GPT-5.6 Sol (max)
Agentic
Tasks the model runs in several steps.
56.2
50.5
Claude Code Claude Opus 5 (Adaptive Reasoning, Max Effort)
Codex GPT-5.6 Sol (max)
These describe the model on its own, not the harness, the tools or the whole setup this page compares — and the lab runs each model at maximum effort, which is not what the composer above is set to. Scores are only comparable within one index version.
Source: Artificial Analysis — their published model index, taken 2026-09-16.
Agent performance on real tasks
Published Terminal-Bench 4.0 runs. Each result names its tested configuration.
Accuracy · Terminal-Bench 4.0
Percent of trials the whole configuration solved.
51.8%
37.3%
Claude Code Claude Code · Claude Opus 5 171 of 330 trials · max effort
Codex Codex · GPT-5.6 Sol 123 of 330 trials · max effort
The whisker is the 95% confidence interval. Two configurations whose whiskers overlap are not separated by this benchmark, however far apart their bars look.
Source: Terminal-Bench — their published runs of a harness and model together, taken 2026-09-17.
Claude Code vs Codex: models, pricing and features
Compare these configurations on the same task, choose the result you prefer, then continue your chosen session. Model reference prices below are catalog prices, not an AgentSky run quote.
| Configuration | Claude Code | Codex |
|---|---|---|
| Model | Claude Opus 5 | GPT-5.6 Sol |
| Thinking level | High | Medium |
| Catalog | Claude Code anthropic/claude-opus-5 | Codex openai/gpt-5.6-sol |
|---|---|---|
| Input$ / 1M tokens | 5.00 | 2.00 |
| Output$ / 1M tokens | 25.00 | 10.00 |
| Context windowtokens | 1M | 1.1M |
| Providers serving it | 5 | 3 |
List price for the model alone. What a run from this page costs also depends on the harness, the tools it calls and how long it works.
Source: OpenRouter — the open model catalog, taken 2026-09-17.