Claude Code vs. Codex

Name: Coding Agent Index
Creator: Artificial Analysis
License: https://artificialanalysis.ai/docs/legal/Terms-of-Use.pdf

Comparação entre Claude Code e Codex no Artificial Analysis Coding Agent Index, incluindo pontuações de benchmarks, custo, tempo de execução e uso de tokens.

Para saber mais sobre nossa metodologia, consulte a página de metodologia.

Explore outras comparações

Coding Agent Index

Artificial Analysis Coding Agent Index v1.5 · Higher is better

Tempo por tarefa

Average agent wall time per task · Lower is better

Custo por tarefa

Average API cost per task (USD) · Lower is better

Comparação

Comparação lado a lado de Claude Code e Codex.

Comparação de agentes de programação

Métrica	Claude Code Sonnet 5.5 (max)	Codex GPT-6.1 Sol (xhigh)	Análise
Harness do agente	Claude Code	Codex
Modelo representativo	Sonnet 5.5 (max)	GPT-6.1 Sol (xhigh)
Coding Agent Index	68	63	Claude Code tem um Coding Agent Index mais alto que Codex
DeepSWE v1.1	72%	73%	Codex tem uma pontuação mais alta em DeepSWE v1.1 que Claude Code
Terminal-Bench 4.0	66%	55%	Claude Code tem uma pontuação mais alta em Terminal-Bench 4.0 que Codex
SWE-Atlas-QnA	67%	61%	Claude Code tem uma pontuação mais alta em SWE-Atlas-QnA que Codex
Custo por tarefa	US$ 14,19	US$ 1,04	Codex tem um custo por tarefa mais baixo que Claude Code
Tempo por tarefa	1,5 h	15,5 min	Codex tem um tempo por tarefa mais baixo que Claude Code
Turnos por tarefa	265,9	40,1	Codex tem menos turnos por tarefa que Claude Code
Uso de tokens por tarefa	27,7 mi	3,2 mi	Codex tem um uso de tokens por tarefa mais baixo que Claude Code
Taxa de acertos de cache	95%	93%	Claude Code tem uma taxa de acertos de cache mais alta que Codex

Variantes de modelo

Variantes de modelo avaliadas para Claude Code e Codex.

Variantes de modelo


Claude Code	Sonnet 5.5 (max)	68	72%	66%	67%	US$ 14,19	1,5 h	27,7 mi
Claude Code	Opus 5.5 (max)	66	68%	63%	66%	US$ 13,04	1,1 h	15,6 mi
Claude Code	Sonnet 5.5 (xhigh)	63	68%	58%	62%	US$ 3,33	27,0 min	7,1 mi
Claude Code	Fable 5.1 (max) (with fallback)	62	64%	58%	65%	US$ 12,39	34,8 min	5,7 mi
Claude Code	Opus 5 (max)	60	63%	55%	62%	US$ 10,79	41,9 min	11,4 mi
Claude Code	Sonnet 5.5 (high)	55	67%	42%	56%	US$ 1,24	12,3 min	2,7 mi
Claude Code	Sonnet 5.5 (medium)	46	65%	27%	45%	US$ 0,62	8,5 min	1,3 mi
Claude Code	Qwen3.8 Max	43	51%	17%	62%	US$ 3,48	1,1 h	8,7 mi
Claude Code	Sonnet 5.5 (low)	42	62%	25%	39%	US$ 0,48	6,3 min	977,3 mil
Codex	GPT-6.1 Sol (xhigh)	63	73%	55%	61%	US$ 1,04	15,5 min	3,2 mi
Codex	GPT-6 Astra (max)	62	68%	56%	62%	US$ 7,47	29,4 min	3,3 mi
Codex	GPT-6.1 Sol (medium)	61	72%	52%	61%	US$ 0,70	10,9 min	2,3 mi
Codex	GPT-6.1 Sol (high)	60	71%	50%	60%	US$ 0,89	13,3 min	2,8 mi
Codex	GPT-6.1 Sol (max)	60	70%	53%	58%	US$ 1,55	24,4 min	4 mi
Codex	GPT-6.1 Sol (low)	57	68%	49%	55%	US$ 0,50	8,6 min	1,7 mi
Codex	GPT-6 Sol (max)	57	69%	43%	58%	US$ 2,99	22,3 min	9,8 mi
Codex	GPT-5.6 Sol (max)	55	72%	37%	54%	US$ 6,35	20,6 min	10,2 mi
Codex	GPT-5.6 Luna (max)	43	66%	15%	49%	US$ 0,44	23,5 min	14,9 mi
Codex	DeepSeek V4 Pro 0813 (max)	43	57%	10%	62%	US$ 0,24	40,3 min	24,2 mi
Codex	GPT-6 Luna (max)	41	64%	15%	44%	US$ 0,18	21,4 min	10,2 mi
Codex	DeepSeek V4 Flash 0731 (max)	39	54%	11%	51%	US$ 0,09	18,3 min	15,4 mi

Desempenho

Desempenho no Artificial Analysis Coding Agent Index.

Artificial Analysis Coding Agent Index

Artificial Analysis Coding Agent Index v1.5 incorporates 3 benchmarks: DeepSWE v1.1, Terminal-Bench 4.0, and SWE-Atlas-QnA · Higher is better

Colorir por

The Artificial Analysis Coding Agent Index is a composite score built from DeepSWE v1.1, Terminal-Bench 4.0 e SWE-Atlas-QnA.

It is useful for quick comparison, but it should be read alongside the per-eval breakdowns. Two agents with similar index values can still have different strengths across repository tasks, terminal workflows, and rubric-based evaluations.

Artificial Analysis Coding Agent Index vs. custo por tarefa

Artificial Analysis Coding Agent Index vs. média do custo da API de pagamento por token por tarefa (USD)

Colorir por

Most attractive quadrant

Pareto line

Each point represents a coding-agent variant. Farther left means lower average cost per task, while higher on the chart means higher benchmark performance. The most efficient agents sit toward the upper-left: stronger results at lower cost.

Uso de tokens

Consumo de tokens no Artificial Analysis Coding Agent Index.

Uso de tokens por tarefa

Média de tokens de entrada, cache e saída por tarefa

As taxas de acerto do cache de prompts podem variar significativamente conforme o roteamento do provedor, alterando materialmente o custo efetivo.

Non-cached input tokens sent to the model, including prompts, instructions, tool context, and task context that were not served from prompt cache.

Artificial Analysis Coding Agent Index vs. tokens totais

Artificial Analysis Coding Agent Index vs. média de tokens totais por tarefa

Colorir por

Most attractive quadrant

Each point represents a coding-agent variant. Farther left means lower average total token usage per task, while higher on the chart means higher benchmark performance. Agents toward the upper-left achieve stronger results with fewer tokens.

Custo

Custo da API de pagamento por token no Artificial Analysis Coding Agent Index, com base nos preços atuais por token.

Custo por tarefa

Average pay-per-token API cost per task (USD) · Lower is better

Colorir por

This chart shows the average pay-per-token API cost per task across the Artificial Analysis Coding Agent Index, spanning DeepSWE v1.1, Terminal-Bench 4.0 e SWE-Atlas-QnA.

Where applicable, that cost model includes standard input pricing, discounted cached-input pricing, separate cache-write charges, and output pricing rather than treating all prompt tokens as if they were billed at the same uncached input rate.

It is intended to show pay-per-token API cost, not consumer plan pricing or the full operational cost of deploying the system in production. Infrastructure, engineering, and supervision costs are not the focus of this metric.

Tempo de execução

Tempo de execução ativo do agente no Artificial Analysis Coding Agent Index.

Tempo por tarefa

Average agent wall time per task · Lower is better

Colorir por

This chart uses agent wall time: how long the agent process was actively running on each task.

It does not include environment startup, verifier or judge time, or other harness overhead, so it is a cleaner comparison of how long the agent itself was working.

Artificial Analysis Coding Agent Index vs. tempo de execução

Artificial Analysis Coding Agent Index vs. média do tempo de execução do agente por tarefa

Colorir por

Most attractive quadrant

Each point represents a coding-agent variant. Farther left means shorter average agent runtime per task, while higher on the chart means higher benchmark performance. Agents toward the upper-left deliver stronger results in less active agent time.

Claude Code vs. Codex

Coding Agent Index

Tempo por tarefa

Custo por tarefa

Comparação

Comparação de agentes de programação

Variantes de modelo

Variantes de modelo

Desempenho

Artificial Analysis Coding Agent Index

What This Metric Means

Artificial Analysis Coding Agent Index vs. custo por tarefa

How to Read This Chart

Uso de tokens

Uso de tokens por tarefa

Input Tokens

Artificial Analysis Coding Agent Index vs. tokens totais

How to Read This Chart

Custo

Custo por tarefa

What Cost Is Measuring

Tempo de execução

Tempo por tarefa

What Execution Time Is Measuring

Artificial Analysis Coding Agent Index vs. tempo de execução

How to Read This Chart