Model comparison
GPT-5.1 vs GPT-5.2
Head-to-head evidence from 16 shared benchmark results across 7 categories. Overall scores shown here use the public BenchAlign v5 ranking lane.
Public leaderboard positions: GPT-5.1 #94 (Estimated); GPT-5.2 #64 (Estimated). Intervals and evidence labels describe ranking uncertainty, not a guarantee for a specific workload.
Evidence parity. GPT-5.1 and GPT-5.2 share 16 comparable benchmark results. 1 of 8 categories are comparable. 5 results are unique to GPT-5.1; 12 to GPT-5.2.
Updated July 20, 2026- Shared results
- 16
- GPT-5.1 only
- 5
- GPT-5.2 only
- 12
- Comparable categories
- 1 / 8
Pick GPT-5.2 if you want the stronger benchmark profile. GPT-5.1 only becomes the better choice if you want the cheaper token bill.
Confidence note. This is a partial-evidence comparison with 16 shared benchmark results across 7 evidence categories; 1 of 8 categories currently have scoreable aggregates for both models. Treat the verdict as directional until coverage is more balanced.
Why this result
GPT-5.2 is clearly ahead on the BenchAlign aggregate, 58.43 to 53.65. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.
GPT-5.2's sharpest advantage is in mathematics, where it averages 35.2 against 26.4. The single biggest benchmark swing on the page is FrontierMath v2 (Tiers 1-3), 31.034% to 40.700%.
GPT-5.2 is also the more expensive model on tokens at $1.75 input / $14.00 output per 1M tokens, versus $1.25 input / $10.00 output per 1M tokens for GPT-5.1. GPT-5.2 gives you the larger context window at 400K, compared with 200K for GPT-5.1.
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-5.1 | Δ | GPT-5.2 |
|---|---|---|---|
| Math | GPT-5.126.4 | Margin→ 8.8 | GPT-5.235.2 |
| Agentic | GPT-5.1Not measured | MarginNo overlap | GPT-5.255.7 |
| Coding | GPT-5.1Not measured | MarginNo overlap | GPT-5.270.6 |
| Reasoning | GPT-5.1Not measured | MarginNo overlap | GPT-5.252.9 |
| Knowledge | GPT-5.1Not measured | MarginNo overlap | GPT-5.292.4 |
| Multimodal | GPT-5.1Not measured | MarginNo overlap | GPT-5.280.4 |
Decisive benchmark drivers
The largest measured benchmark gaps in this matchup, with exact reported values.
More
- Source ↗
FrontierMath v2 (Tiers 1-3)
MathA 31.034%B 40.700%Winner: GPT-5.2Δ 9.7FrontierMath v2 (Tiers 1-3): GPT-5.1 scored 31.034%; GPT-5.2 scored 40.700%. GPT-5.2 wins this benchmark. - Source ↗
FrontierMath v2 (Tier 4)
MathA 12.500%B 18.800%Winner: GPT-5.2Δ 6.3FrontierMath v2 (Tier 4): GPT-5.1 scored 12.500%; GPT-5.2 scored 18.800%. GPT-5.2 wins this benchmark.
Operational comparison
Runtime and commercial metrics are compared only when both models have a complete sourced value.
| Metric | GPT-5.1 | GPT-5.2 | Comparison |
|---|---|---|---|
| Input / output priceUSD per 1M tokens | GPT-5.1$1.25 input / $10 output | GPT-5.2$1.75 input / $14 output | GPT-5.1 has the lower combined listed price. |
| Generation speedtokens per second | GPT-5.1111 tok/s | GPT-5.273 tok/s | GPT-5.1 has the higher measured throughput. |
| First-answer latencyseconds to first token | GPT-5.157.47 s | GPT-5.2130.34 s | GPT-5.1 reaches the first token sooner. |
| Context windowmaximum listed tokens | GPT-5.1200K | GPT-5.2400K | GPT-5.2 lists the larger context window. |
Benchmark Deep Dive
Agentic8 benchmarks
| Benchmark | GPT-5.1 | GPT-5.2 | Result |
|---|---|---|---|
| AA Agentic IndexSource | 21.0% | — | Not comparable |
| τ²-bench resultsSource | 81.9% | 84.8% | GPT-5.2 leads |
| Gert LabsSource | 41.24% | 46.54% | GPT-5.2 leads |
| GDPval-AASource | 24.4% | — | Not comparable |
| GDPval-AASource | 987 | — | Not comparable |
| BrowseCompSource | — | 65.8% | Not comparable |
| OSWorld-VerifiedSource | — | 47.3% | Not comparable |
| JobBenchSource | — | 34.3% | Not comparable |
Coding5 benchmarks
Reasoning3 benchmarks
Knowledge7 benchmarks
| Benchmark | GPT-5.1 | GPT-5.2 | Result |
|---|---|---|---|
| Artificial Analysis Intelligence IndexSource | 36.9% | 42.2% | GPT-5.2 leads |
| AA-GPQA DiamondSource | 87.3% | 90.3% | GPT-5.2 leads |
| AA-HLESource | 26.5% | 35.4% | GPT-5.2 leads |
| AA-Omniscience IndexSource | 5.6% | -1.0% | GPT-5.1 leads |
| AA-Omniscience AccuracySource | 37.6% | 43.8% | GPT-5.2 leads |
| AA-Omniscience Hallucination RateSource | 51.3% | 79.7% | GPT-5.1 leads |
| GPQASource | — | 92.4% | Not comparable |
MathGPT-5.2 wins3 benchmarks
Multimodal6 benchmarks
Inst. Following1 benchmarks
| Benchmark | GPT-5.1 | GPT-5.2 | Result |
|---|---|---|---|
| AA-IFBenchSource | 72.9% | 75.4% | GPT-5.2 leads |
Frequently Asked Questions (2)
Which is better, GPT-5.1 or GPT-5.2?
GPT-5.2 is ahead on BenchLM's BenchAlign leaderboard, 58.43 to 53.65. The biggest single separator in this matchup is FrontierMath v2 (Tiers 1-3), where the scores are 31.034% and 40.700%.
Which is better for math, GPT-5.1 or GPT-5.2?
GPT-5.2 has the edge for math in this comparison, averaging 35.2 versus 26.4. Inside this category, FrontierMath v2 (Tiers 1-3) 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.