GPT-5.2 ProvsOpus 4.5
Across 6 shared benchmarks, GPT-5.2 Pro leads overall: GPT-5.2 Pro wins 5, Opus 4.5 wins 1, with 0 ties and an average score difference of +10.43.
GPT-5.2 Pro
OpenAI · 2025-12-11 · Reasoning model
Opus 4.5
Anthropic · 2025-11-25 · Reasoning model
GPT-5.2 Pro5 wins(83%)(17%)1 winOpus 4.5
Benchmark scores
Grouped by capability, sorted by largest gap within each. 6 shared benchmarks.
General Knowledge
GPT-5.2 Pro 4/4| Benchmark | GPT-5.2 Pro | Opus 4.5 | Diff |
|---|---|---|---|
| ARC-AGI-2 | 54.2023 / 62 | 37.6029 / 62Extended (no tools) | +16.60 |
| ARC-AGI | 90.5017 / 68 | 8024 / 68Extended (no tools) | +10.50 |
| HLE | 5029 / 172 | 43.2049 / 172Extended (with tools) | +6.80 |
| GPQA Diamond | 93.209 / 187 | 8742 / 187Extended (no tools) | +6.20 |
Commonsense Reasoning
Opus 4.5 1/1| Benchmark | GPT-5.2 Pro | Opus 4.5 | Diff |
|---|---|---|---|
| Simple Bench | 57.4019 / 63极高强度思考（无工具） | 6212 / 63Extended (no tools) | -4.60 |
Math and Reasoning
GPT-5.2 Pro 1/1| Benchmark | GPT-5.2 Pro | Opus 4.5 | Diff |
|---|---|---|---|
| FrontierMath - Tier 4 | 31.309 / 80 | 4.2040 / 80Normal (No Tools) | +27.10 |
Specs
| Field | GPT-5.2 Pro | Opus 4.5 |
|---|---|---|
| Publisher | OpenAI | Anthropic |
| Release date | 2025-12-11 | 2025-11-25 |
| Model type | Reasoning model | Reasoning model |
| Architecture | Dense | Dense |
| Parameters | Not available | Not available |
| Context length | 256K | 200K |
| Max output | Not available | 64K |
API pricing
Prices use DataLearner records when available; missing fields are not inferred.
| Item | GPT-5.2 Pro | Opus 4.5 |
|---|---|---|
| Text input | $21 / 1M tokens | $5 / 1M tokens |
| Text output | $168 / 1M tokens | $25 / 1M tokens |
| Cache read | Not public | $0.5 / 1M tokens |
| Cache write | Not public | $6.25 / 1M tokens |
Summary
- GPT-5.2 Proleads in:General Knowledge (4/4), Math and Reasoning (1/1)
- Opus 4.5leads in:Commonsense Reasoning (1/1)
On average across the 6 shared benchmarks, GPT-5.2 Pro scores 10.43 higher.
Largest single-benchmark gap: FrontierMath - Tier 4 — GPT-5.2 Pro 31.30 vs Opus 4.5 4.20 (+27.10).
Page generated from structured model, pricing and benchmark records. No real-time LLM is used to write the prose.