Head to head
GPT-4o vs GPT-4.1
GPT-4o from OpenAI against GPT-4.1 from OpenAI — specification, price and every benchmark both makers have published, in one table.
Benchmarks
GPT-4.1 leads
GPT-4.1 wins 3 of the 3 benchmarks both models report, GPT-4o wins 0, by a wide margin. The average gap across shared tests is 8.1 points.
Price
GPT-4.1 is cheaper
On a 3:1 input-to-output mix, GPT-4.1 costs $3.50 per million tokens against $4.38 for GPT-4o — about 1.2× less.
What actually differs
- GPT-4.1 takes 1M tokens of context against 128K — 8.2× more room for long documents or a large codebase.
Scorecard
Which is better at what
Maths, coding, reasoning and the rest — one line each, averaged over the benchmarks both models actually report.
| Category | GPT-4o | GPT-4.1 | Better at this |
|---|---|---|---|
| ReasoningGPQA Diamond | 53.6% | 66.3% | GPT-4.1+12.7 |
| MathsNeither model reports this | — | — | Not comparable |
| CodingOnly one model reports this | — | — | Not comparable |
| KnowledgeMMLU-Pro | 74.7% | 80.5% | GPT-4.1+5.8 |
| MultimodalMMMU | 69.1% | 74.8% | GPT-4.1+5.7 |
| Instruction followingOnly one model reports this | — | — | Not comparable |
| Human preferenceOnly one model reports this | — | — | Not comparable |
| Categories wonOut of 3 comparable | 0 | 3 | GPT-4.1 |
Each category averages only the benchmarks every model here reports, so no one is credited for a test the other did not run. A category with no shared test is marked Not comparable rather than guessed at.
Side by side
GPT-4o and GPT-4.1, row by row
| Attribute | GPT-4oOpenAI | GPT-4.1OpenAI |
|---|---|---|
| Specification | ||
| MakerWho built it | OpenAI | OpenAI |
| Released | 2024-05 | 2025-04 |
| Context windowHow much can go in at once | 128,000 tokens | 1,047,576 tokens |
| Max output | 16,384 tokens | 32,768 tokens |
| Input | Text, Image, Audio | Text, Image |
| ReasoningSpends extra tokens thinking before it answers | No | No |
| Tool calling | Yes | Yes |
| Knowledge cutoff | 2023-10 | 2024-06 |
| Licence | Proprietary | Proprietary |
| Open weightsCan you download and run it yourself | No | No |
| Price | ||
| Input priceUSD per million tokens in | $2.50 | $2.00 |
| Output priceUSD per million tokens out | $10.00 | $8.00 |
| Cached input | $1.25 | $0.50 |
| Blended 3:1A 3-in-to-1-out million-token mix — a fairer single number than input price alone | $4.38 | $3.50Cheapest |
| Published benchmarks | ||
| MMLU-Pro12,000 reasoning-heavy multiple-choice questions across 14 academic subjects, with ten options instead of four. The harder successor to MMLU. | ||
| GPQA Diamond198 graduate-level physics, chemistry and biology questions written to be Google-proof. PhD holders in the matching field score about 65%. | ||
| SWE-bench Verified500 human-validated GitHub issues from real Python repositories. The model must produce a patch that makes the project's own tests pass. | Not reported | |
| HumanEval164 short Python functions written from a docstring. Saturated at the frontier — kept here for continuity with older models. | Not reported | |
| MMMUCollege-level questions that require reading charts, diagrams, tables and photographs alongside the text. | ||
| IFEvalVerifiable instructions — word counts, formats, forbidden words — checked by a program rather than a judge model. | Not reported | |
| LMArena EloElo rating from blind pairwise votes by the public on LMArena. Measures what people prefer, not what is correct. | Not reported | |
| Links | Full page | Full page |
| Row verified | 2026-08 | 2026-08 |
Where these numbers come from
Every score on this page is a published figure, taken from the model's own card, system card, technical report or release post, or from a public leaderboard. CorX Labs did not run these evaluations. Most are self-reported by the lab that built the model, which means they were produced under that lab's own choice of prompt, scaffold and number of attempts — so treat them as a starting point for a shortlist, not as a settled ranking.
A score someone other than the model's maker measured is marked Independent and names its measurer. Those are the stronger numbers on this page — an outside harness has no reason to flatter anyone — and there are not many of them.
Where a figure has not been published, the cell reads Not reported rather than an estimate. Nothing here is inferred, interpolated or guessed. Each model records the month its row was last checked. Full method and caveats.
Questions
GPT-4o or GPT-4.1?
Is GPT-4o better than GPT-4.1?
GPT-4.1 wins 3 of the 3 benchmarks both models report, GPT-4o wins 0, by a wide margin. The average gap across shared tests is 8.1 points.
Which is cheaper, GPT-4o or GPT-4.1?
On a 3:1 input-to-output mix, GPT-4.1 costs $3.50 per million tokens against $4.38 for GPT-4o — about 1.2× less.
What is the difference between GPT-4o and GPT-4.1?
GPT-4.1 takes 1M tokens of context against 128K — 8.2× more room for long documents or a large codebase.
