Model comparison
GPT-5.2 vs GPT-5.3 Codex
Head-to-head evidence from 18 shared benchmark results across 6 categories. Overall scores shown here use the public BenchAlign v5 ranking lane.
Public leaderboard positions: GPT-5.2 #64 (Estimated); GPT-5.3 Codex #26 (Supported). Intervals and evidence labels describe ranking uncertainty, not a guarantee for a specific workload.
Evidence parity. GPT-5.2 and GPT-5.3 Codex share 18 comparable benchmark results. 2 of 8 categories are comparable. 10 results are unique to GPT-5.2; 3 to GPT-5.3 Codex.
Updated July 20, 2026- Shared results
- 18
- GPT-5.2 only
- 10
- GPT-5.3 Codex only
- 3
- Comparable categories
- 2 / 8
Pick GPT-5.3 Codex if you want the stronger benchmark profile. GPT-5.2 only becomes the better choice if coding is the priority.
Confidence note. This is a partial-evidence comparison with 18 shared benchmark results across 6 evidence categories; 2 of 8 categories currently have scoreable aggregates for both models. Treat the verdict as directional until coverage is more balanced.
Why this result
GPT-5.3 Codex is clearly ahead on the BenchAlign aggregate, 66.69 to 58.43. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.
GPT-5.3 Codex's sharpest advantage is in agentic, where it averages 71.4 against 55.7. The single biggest benchmark swing on the page is OSWorld-Verified, 47.3% to 64.7%. GPT-5.2 does hit back in coding, so the answer changes if that is the part of the workload you care about most.
Category breakdown
Exact category averages are shown below. Not measured means BenchLM does not have enough sourced public coverage for that model and category.
| Category | GPT-5.2 | Δ | GPT-5.3 Codex |
|---|---|---|---|
| Agentic | GPT-5.255.7 | Margin→ 15.7 | GPT-5.3 Codex71.4 |
| Coding | GPT-5.270.6 | Margin← 3.4 | GPT-5.3 Codex67.2 |
| Reasoning | GPT-5.252.9 | MarginNo overlap | GPT-5.3 CodexNot measured |
| Knowledge | GPT-5.292.4 | MarginNo overlap | GPT-5.3 CodexNot measured |
| Math | GPT-5.235.2 | MarginNo overlap | GPT-5.3 CodexNot measured |
| Multimodal | GPT-5.280.4 | MarginNo overlap | GPT-5.3 CodexNot measured |
Decisive benchmark drivers
The largest measured benchmark gaps in this matchup, with exact reported values.
More
- Source ↗
OSWorld-Verified
AgenticA 47.3%B 64.7%Winner: GPT-5.3 CodexΔ 17.4OSWorld-Verified: GPT-5.2 scored 47.3%; GPT-5.3 Codex scored 64.7%. GPT-5.3 Codex wins this benchmark. - Source ↗
SWE-bench Verified
CodingA 80%B 85%Winner: GPT-5.3 CodexΔ 5SWE-bench Verified: GPT-5.2 scored 80%; GPT-5.3 Codex scored 85%. GPT-5.3 Codex wins this benchmark. - Source ↗
SWE-bench Pro
CodingA 55.6%B 56.8%Winner: GPT-5.3 CodexΔ 1.2SWE-bench Pro: GPT-5.2 scored 55.6%; GPT-5.3 Codex scored 56.8%. GPT-5.3 Codex wins this benchmark.
Operational comparison
Runtime and commercial metrics are compared only when both models have a complete sourced value.
| Metric | GPT-5.2 | GPT-5.3 Codex | Comparison |
|---|---|---|---|
| Input / output priceUSD per 1M tokens | GPT-5.2$1.75 input / $14 output | GPT-5.3 Codex$1.75 input / $14 output | Listed prices are equal. |
| Generation speedtokens per second | GPT-5.273 tok/s | GPT-5.3 Codex79 tok/s | GPT-5.3 Codex has the higher measured throughput. |
| First-answer latencyseconds to first token | GPT-5.2130.34 s | GPT-5.3 Codex88.26 s | GPT-5.3 Codex reaches the first token sooner. |
| Context windowmaximum listed tokens | GPT-5.2400K | GPT-5.3 Codex400K | Listed context windows are equal. |
Benchmark Deep Dive
AgenticGPT-5.3 Codex wins6 benchmarks
| Benchmark | GPT-5.2 | GPT-5.3 Codex | Result |
|---|---|---|---|
| BrowseCompSource | 65.8% | — | Not comparable |
| OSWorld-VerifiedSource | 47.3% | 64.7% | GPT-5.3 Codex leads |
| τ²-bench resultsSource | 84.8% | 86% | GPT-5.3 Codex leads |
| Gert LabsSource | 46.54% | 57.47% | GPT-5.3 Codex leads |
| JobBenchSource | 34.3% | 33.7% | GPT-5.2 leads |
| Terminal-Bench 2.0Source | — | 77.3% | Not comparable |
CodingGPT-5.2 wins5 benchmarks
Reasoning3 benchmarks
Knowledge7 benchmarks
| Benchmark | GPT-5.2 | GPT-5.3 Codex | Result |
|---|---|---|---|
| GPQASource | 92.4% | — | Not comparable |
| Artificial Analysis Intelligence IndexSource | 42.2% | 44.3% | GPT-5.3 Codex leads |
| AA-GPQA DiamondSource | 90.3% | 91.5% | GPT-5.3 Codex leads |
| AA-HLESource | 35.4% | 39.9% | GPT-5.3 Codex leads |
| AA-Omniscience IndexSource | -1.0% | 9.9% | GPT-5.3 Codex leads |
| AA-Omniscience AccuracySource | 43.8% | 51.8% | GPT-5.3 Codex leads |
| AA-Omniscience Hallucination RateSource | 79.7% | 86.9% | GPT-5.2 leads |
Math3 benchmarks
Multimodal6 benchmarks
Inst. Following1 benchmarks
| Benchmark | GPT-5.2 | GPT-5.3 Codex | Result |
|---|---|---|---|
| AA-IFBenchSource | 75.4% | 75.4% | Tie |
Frequently Asked Questions (3)
Which is better, GPT-5.2 or GPT-5.3 Codex?
GPT-5.3 Codex is ahead on BenchLM's BenchAlign leaderboard, 66.69 to 58.43. The biggest single separator in this matchup is OSWorld-Verified, where the scores are 47.3% and 64.7%.
Which is better for coding, GPT-5.2 or GPT-5.3 Codex?
GPT-5.2 has the edge for coding in this comparison, averaging 70.6 versus 67.2. Inside this category, Vibe Code Bench is the benchmark that creates the most daylight between them.
Which is better for agentic tasks, GPT-5.2 or GPT-5.3 Codex?
GPT-5.3 Codex has the edge for agentic tasks in this comparison, averaging 71.4 versus 55.7. Inside this category, OSWorld-Verified is the benchmark that creates the most daylight between them.
Related Comparisons
Explore More
Choose a model with this week’s evidence
Join 2,000+ readers for ranking moves, pricing changes, and the claims that still need proof.
One email each week. Unsubscribe anytime.