← All Comparisons
Claude Sonnet 4.6 vs GPT-5.2 Codex
A detailed comparison of Claude Sonnet 4.6 (Anthropic) and GPT-5.2 Codex (OpenAI) across pricing, performance, and features.
Pricing Comparison
| Metric | Claude Sonnet 4.6 | GPT-5.2 Codex | Difference |
|---|---|---|---|
| Input / 1M tokens | $3.00 | $1.75 | -42% |
| Output / 1M tokens | $15.00 | $14.00 | -7% |
| Context window | 200K | 200K | — |
| Max output | 16K | 65.536K | — |
Benchmark Comparison
| Benchmark | Claude Sonnet 4.6 | GPT-5.2 Codex |
|---|---|---|
| MMLU-Pro | 86% | 89% |
| HumanEval | 94% | 95.5% |
| GPQA | 70% | 76% |
Capabilities
| Capability | Claude Sonnet 4.6 | GPT-5.2 Codex |
|---|---|---|
| code | ✓ | ✓ |
| reasoning | ✓ | ✓ |
| text | ✓ | ✓ |
| tool-use | ✓ | ✓ |
| vision | ✓ | ✓ |
Claude Sonnet 4.6 Strengths
- ✓Opus 4.5 quality at 1/5th the cost
- ✓Best value for production workloads
- ✓1M context in beta
Claude Sonnet 4.6 Weaknesses
- ✗Long context pricing doubles above 200K
- ✗Slightly below Opus 4.6 on hardest tasks
GPT-5.2 Codex Strengths
- ✓Strong coding performance
- ✓Available via API now
- ✓200K context window
GPT-5.2 Codex Weaknesses
- ✗Being superseded by 5.3
- ✗Higher output cost than some competitors
Quick Verdict
Best value: GPT-5.2 Codex is the more affordable option at $1.75/$14 per 1M tokens.
Higher benchmarks: GPT-5.2 Codex scores higher on average across available benchmarks (86.8% avg).
Choose GPT-5.2 Codex if cost matters most. Choose Claude Sonnet 4.6 if you need the best possible quality for complex tasks.