Compare / head-to-head
GPT-4.1vso3
o3 leads 12 of 12 shared benchmarks. Both list the same combined token price. GPT-4.1 has the larger context window (1.0M tokens).
Benchmarks from cited public sources; pricing from official pages; status from official provider feeds.
Shared benchmarks
0 – 12
o3 leads
Cheaper per token
Tie
list price, input + output
Larger context
GPT-4.1
1.0M tokens
Providers
OpenAI
same provider
GPT-4.1
OpenAI · released 2025-04-14
textvision
Quality
Benchmark matrix
12 shared · 3 only GPT-4.1 · 5 only o3| Benchmark | GPT-4.1 | o3 | Δ | Edge |
|---|---|---|---|---|
| Reported by both · 12 | ||||
| SWE-bench Verified | 54.6% ↗ | 69.1% ↗ | -14.5 pt | o3 |
| GPQA Diamond | 66.3% ↗ | 81.8% ↗epoch run | -15.5 pt | o3 |
| LMArena Elo | 1414.8 ↗ | 1431.6 ↗ | -16.8 | o3 |
| Chess Puzzles (Epoch AI run) | 6% ↗epoch run | 38% ↗epoch run | -32 pt | o3 |
| FrontierMath Tiers 1-3 v2 (Epoch AI run) | 6% ↗epoch run | 33.3% ↗epoch run | -27.3 pt | o3 |
| LMArena Vision (LMArena) | 1214 ↗ | 1215.7 ↗ | -1.7 | o3 |
| MATH Level 5 (Epoch AI run) | 83% ↗epoch run | 97.8% ↗epoch run | -14.8 pt | o3 |
| OTIS Mock AIME 2024-2025 (Epoch AI run) | 38.3% ↗epoch run | 84.4% ↗epoch run | -46.1 pt | o3 |
| SimpleBench (SimpleBench) | 27% ↗ | 53.1% ↗ | -26.1 pt | o3 |
| SimpleQA Verified | 31.1% ↗epoch run | 49.4% ↗epoch run | -18.3 pt | o3 |
| SWE-bench Verified (Epoch AI run) | 48.5% ↗epoch run | 62.3% ↗epoch run | -13.8 pt | o3 |
| WeirdML (Håvard Tveit Ihle) | 39% ↗ | 52.4% ↗ | -13.4 pt | o3 |
| Only GPT-4.1 reports · 3 | ||||
| MMLU | 90.2% ↗ | not reported | — | — |
| AIME 2024 | 48.1% ↗ | not reported | — | — |
| Video-MME (long, no subtitles) | 72% ↗ | not reported | — | — |
| Only o3 reports · 5 | ||||
| AIME 2025 | not reported | 89.2% ↗matharena ⚠ | — | — |
| GSO Opt@1 (GSO) | not reported | 3.9% ↗ | — | — |
| Mystery Game Puzzles (Epoch AI run) | not reported | 29% ↗epoch run | — | — |
| OSWorld-Verified (XLANG) | not reported | 23% ↗ | — | — |
| SAGE (Vals AI) | not reported | 41.8% ↗ | — | — |
Scores tagged "epoch" or "matharena" are independent runs, used only where the lab has not published its own; ⚠ marks rows MathArena flags as released after the competition. Higher is better on every row. Δ is GPT-4.1 minus o3 in the benchmark's own unit. "Not reported" means the lab has not published that figure; it is not a zero. ↗ opens the source.
Specs & pricing
Side by side
Official pricing pages and model cards| Spec | GPT-4.1 | o3 | Edge |
|---|---|---|---|
| Context window | 1.0M tokens | 200K tokens | GPT-4.1 |
| Max output | 33K tokens | 100K tokens | o3 |
| Input price / 1M | $2 ↗ | $2 ↗ | Tie |
| Output price / 1M | $8 ↗ | $8 ↗ | Tie |
| Cached input / 1M | $0.5 ↗ | $0.5 ↗ | Tie |
| Input + output / 1M Lower is cheaper. List prices; batch, tool and regional fees excluded. | $10 | $10 | Tie |
| Modalities | text · vision | text · vision | Tie |
| Released | 2025-04-14 | 2025-04-16 | — |
| Cited benchmark scores | 21 | 22 | — |
Reliability
OpenAI status
Live from /statusMore matchups
GPT-4.1 vs …
Models sharing the most benchmarksMore matchups
o3 vs …
Models sharing the most benchmarksBuilt by Respan
Which one wins on your data?
Public benchmarks are a starting point. Run GPT-4.1 and o3 on your own prompts with Respan evals, or route to either through one gateway key with automatic failover.