메인 콘텐츠로 건너뛰기
OpenAI

2025년 4월 16일

릴리스제품

OpenAI o3 및 o4-mini를 소개합니다

모든 도구를 완전히 활용할 수 있는 현재 가장 스마트하고 유능한 모델

로딩 중...

2025년 6월 10일 업데이트: Pro 사용자는 이제 ChatGPT와 우리 API에서 OpenAI o3‑pro를 사용할 수 있습니다.. OpenAI o1‑pro와 같이, o3‑pro는 가장 똑똑한 모델 OpenAI o3의 버전 중 하나로 더 길게 생각하고 가장 신뢰할 수 있는 응답을 제공하도록 설계되었습니다. 릴리스 노트(새 창에서 열기)에서 전체 내용을 확인하실 수 있습니다.


오늘 자로, 응답하기 전에 더 오랜 시간 생각하도록 훈련된 o-시리즈의 최신 모델, OpenAI o3o4-mini를 출시합니다. 지금까지 출시한 모델 중 가장 스마트한 모델로, 궁금한 것이 많은 사용자부터 심층적인 내용이 필요한 연구자까지 모두를 위한 ChatGPT의 기능이 변화하는 단계를 보여줍니다. 우리의 추론 모델은 최초로 ChatGPT 내에서 모든 도구를 에이전트 방식으로 사용하고 결합할 수 있습니다. 여기에는 웹 검색, 업로드된 파일 및 Python의 다른 데이터 분석, 시각적 입력에 관한 심층 사고뿐 아니라 이미지 생성까지 포함됩니다. 무엇보다도 이러한 모델은 알맞은 출력 형식으로 상세하고 신중한 답변을 생산하기 위해 도구를 언제 어떻게 사용할 것인지 추론하도록 훈련되었으며, 일반적으로 1분 이내에 답변을 제공하고 훨씬 더 복잡한 문제를 처리할 수 있습니다. 이를 통해 다면적인 문제를 더욱 효율적으로 처리할 수 있으며, 사용자를 대신해 독립적으로 작업을 수행할 수 있는 에이전트 방식의 ChatGPT에 더 가까워짐을 의미합니다. 최첨단 추론과 모든 도구에 대한 액세스를 결합한 강력함은 학술 벤치마크 및 실제 작업에서 특히 더 뛰어난 성능을 드러내며, 지능과 유용성 모든 부분에서 새로운 기준을 제시합니다.

변경 내용

OpenAI o3코딩, 수학, 과학, 시각 인식 등을 포함한 다양한 분야에서 최전선을 개척하는 가장 강력한 추론 모델입니다. Codeforces, SWE-bench(별도의 맞춤형 모델별 스캐폴드 구축 없이), MMMU를 포함한 벤치마크에서 새로운 SOTA 기준을 세웠습니다. 이 모델은 다면적인 분석이 필요하고 명확한 답변이 즉각적으로 도출되지 않는 복잡한 쿼리에 이상적입니다. 특히 이미지, 차트, 그래픽 분석과 같은 시각적 업무에 강력한 성능을 보여줍니다. 외부 전문가의 평가에서 o3는 복잡한 실제 작업에 대해 OpenAI o1보다 주요 에러가 20% 더 적게 발생했으며, 특히 프로그래밍, 비즈니스/컨설팅 및 창의적인 구상에서 뛰어난 성능을 발휘했습니다. 초기 테스트 담당자들은 사고 파트너로서 모델의 분석에 대한 엄격함이 뛰어나다고 평가했으며, 특히 생물학, 수학 및 엔지니어링 컨텍스트에서 새로운 가설을 생성하고 엄격하게 평가하는 기능을 강조했습니다.

OpenAI o4-mini는 빠르고 비용 효율적 사고에 최적화된 소형 모델로, 특히 수학, 코딩 및 시각적 작업에서 규모 및 사이즈에 비해 뛰어난 성능을 발휘합니다. 이 모델은 AIME 2024 및 2025에서 최고 성능을 기록했습니다. 컴퓨터 활용으로 AIME 시험의 난도가 많이 감소하긴 했지만, o4-mini가 Python 인터프리터에 액세스하여 AIME 2025에서 99.5% pass@1(100% consensus@8)를 달성했다는 점은 주목할 만합니다. 이러한 결과를 도구 액세스가 불가능한 모델의 성능과 비교할 수는 없지만, o4-mini가 도구를 효과적으로 사용하는 것을 보여주는 하나의 예시가 될 수 있으며, o3도 AIME 2025에서 도구 사용에 대해 비슷한 성과(98.4% pass@1, 100% consensus@8)를 보여주었습니다.

전문가 평가에서 o4-mini는 non-STEM뿐 아니라 데이터 과학 등과 같은 영역에서 이전 버전인 o3‑mini보다 뛰어난 성능을 발휘하기도 했습니다. 효율성 덕분에 o4-mini는 사용량 제한이 o3보다 월등히 높으며, 추론이 필요한 질문에 대해 높은 처리량을 갖춘 고용량 옵션을 제공하는 모델이 될 수 있습니다. 외부 전문 평가자들은 개선된 지능과 웹 소스 포용성으로 인해 두 모델이 이전 모델과 비교해 지침을 따르는 수준이 향상되고 더 유용하고 입증 가능한 응답을 제공한다고 평가했습니다. 기존에 출시되었던 자사 추론 모델과 비교하면 이 두 모델은 보다 더 맞춤화되고 관련성 높은 응답을 제공하고자 메모리와 과거 대화를 참고하므로, 더욱 자연스럽고 대화에 능한 것으로 느껴질 겁니다.

멀티모달

코딩

All SWE-bench evaluation runs use a fixed subset of n=477 verified tasks which have been validated on our internal infrastructure.

지시를 따르며 에이전틱한 도구 사용

모든 모델은 높은 ‘추론 노력’ 설정에서 평가되었으며, 이는 ChatGPT의 ‘o4-mini-high’ 등과 유사한 수준입니다.

강화 학습으로 꾸준한 확장

OpenAI o3를 개발하는 동안 GPT‑시리즈 사전 훈련에서 관찰된 것과 동일하게, 대규모 강화 학습에서도 “더 많은 연산 = 더 나은 성능” 경향이 나타나는 것을 확인했습니다. 강화 학습의 스케일링 경로를 다시 따라가면서 학습 연산량과 추론 시 추론 능력(Inference-time reasoning) 모두 그 규모를 더 크게 늘렸음에도, 성능 향상이 뚜렷하게 나타났으며 이는 모델이 더 깊이 사고할수록 성능이 계속 향상한다는 것을 입증합니다. OpenAI o1, o3과 동일한 레이턴시 및 비용으로 ChatGPT에서 더욱 뛰어난 성능을 제공합니다. 또한 더 오래 생각할 수록 성능이 향상된다는 점도 입증되었습니다.

이에 더해 두 모델 모두 강화 학습을 통해 도구를 사용하도록 훈련했으며, 단순히 도구를 사용하는 것뿐 아니라 언제 사용해야 하는지 사고하도록 훈련했습니다. 필요한 결과에 따라 도구를 효율적으로 사용하는 기능은 개방형 상황, 특히 시각적 추론과 다단계 워크플로 등에서 더욱 뛰어난 성능을 발휘합니다. 초기 테스트 담당자가 확인한 대로 학술적 벤치마크 및 실제 업무 모두에 이러한 개선 사항이 반영됩니다.

이미지를 이용한 생각

이미지를 이용해 생각하는 ChatGPT

이번 두 모델들은 최초로 사고의 사슬(CoT)에 이미지를 직접 결합할 수 있습니다. 단순이 이미지를 보는 것이 아니라 이미지를 이용해 생각합니다. 이는 시각적 추론과 텍스트 추론을 결합한 새로운 종류의 문제 해결 방식을 가능하게 하며, 멀티모달 벤치마크에서 이러한 최첨단 성능을 확인할 수 있습니다.

사람들은 화이트보드 사진, 교과서 다이어그램 또는 손으로 그린 스케치의 사진 등을 업로드할 수 있고 모델인 이를 해석할 수 있습니다. 이미지가 흐리거나 거꾸로 되어 있거나 품질이 낮아도 문제없습니다. 도구의 사용으로 모델은 상황에 따라 이미지를 처리할 수 있습니다. 회전하거나 확대/축소하거나 추론 과정의 일부로 변환할 수도 있습니다.

이들 모델은 시각적 인식 작업에서 최고 수준의 정확도를 보여주며, 이전에는 처리할 수 있는 영역 밖의 일이었던 문제를 해결할 수 있게 되었습니다. 시각적 추론 리서치 블로그에서 자세한 내용을 확인해 보세요.

에이전틱한 도구 사용

OpenAI o3 및 o4-mini는 ChatGPT 내 모든 도구에 액세스할 수 있을 뿐 아니라 API를 호출하는 기능을 통해 사용자의 맞춤형 도구에도 액세스할 수 있습니다. 이러한 모델은 알맞은 출력 형식으로 상세하고 신중한 답변을 생산하기 위해 도구를 언제 어떻게 사용할 것인지 선택하여 문제를 해결하는 방식을 추론하도록 훈련되었으며, 일반적으로 1분 이내에 답변을 제공할 수 있습니다.

예를 들면 사용자가 다음과 같이 질문합니다. “작년과 비교해서 캘리포니아의 여름 에너지 사용량은 얼마나 될까?” 모델은 다양한 도구를 한꺼번에 호출하여 공개적으로 이용할 수 있는 데이터를 웹에서 검색하고, 예측을 위해 Python 코드를 작성하고, 그래프나 이미지를 생성하고, 예측의 근거가 되는 주요 요인을 설명할 수 있습니다. 추론을 통해 모델은 접한 정보에 따라 필요한만큼 반응하고 방향을 바꿀 수 있습니다. 예를 들어, 검색 공급자의 도움으로 여러번 웹을 검색하여 결과를 찾고, 더 정보가 필요하면 새로운 검색을 시도할 수 있습니다.

이렇게 유연하고 전략적인 접근 방식은 내장된 지식 이상의 최신 정보에 대한 액세스, 확장된 추론, 종합, 다양한 양식으로의 출력 생성 등이 필요한 작업을 모델이 처리할 수 있게 해줍니다.

모든 예시는 OpenAI o3로 완성했습니다.

OpenAI o3

Solve the following very hard math problem using your in-built knowledge of the literature:

Construct a degree 19 polynomial \ (p(x)) \ in \ mathbb{C}[x] such that \ (X := \ {\, p(x) = p(y) \, \} \ subset \ mathbb{P}^{1} \ times \ mathbb{P}^{1}\ ) has at least 3 (but not all linear) irreducible components over \ mathbb{C}. Choose \ (p(x)\ ) to be odd, monic, have real coefficients and linear coefficient \ (-19\ ) and calculate \ (p(19)\ ).

다항식 구성에 대해 55초 동안 이성 사용함
  • The curve defined by p(x) = p(y) factors if p(x) is decomposable or has repeated factors. Actually, p(x)-p(y) always factors, and it’s divisible by (x-y). This gives one factor, and the rest is a degree d-1 polynomial in x and y. For the curve to have at least three irreducible components, p(x)-p(y) must factor into at least three nontrivial factors. Each factor corresponds to a component in projective space, and at least one must be nonlinear.
  • I’m checking the Dickson polynomials D_n(x,a) and focusing on their coefficients. With a=1, it seems the leading coefficient is always 1, confirming that Dickson polynomials are monic. The linear coefficient for odd n is following a predictable sign pattern: for n=19, it gives a coefficient of -19, which is exactly what I need. So the Dickson polynomial D_n(x,1) with odd n fits nicely with the linear coefficient I was looking for.
  • I've confirmed that the linear coefficient for p(x)=D19(x)p(x) = D_{19}(x)p(x)=D19​(x) is correct, with the necessary scaling factor to match. Now for the next step, I need to compute D19(19)D_{19}(19)D19​(19). Since Tn(z)T_n(z)Tn​(z) grows quickly for real z>1z > 1z>1, I plan to compute this value using Python, leveraging the recurrence to avoid mistakes—it's going to be a big number. I'll make sure to present the numeric result neatly so it's understandable since the coefficients are integers.
  • The Dickson polynomial D_{19}(x,1) is confirmed to have the required properties: monic, odd, real coefficients, and a linear coefficient of -19. The factorization of D_{19}(x) into irreducible components is also verified and satisfies the condition of having at least 10 components (1 linear + 9 quadratics). To compute p(19), we'll use the recurrence relation for Dickson polynomials, but there's also a closed-form expression involving cosine that could be handy here. I'll proceed accordingly!
  • I confirmed the linear coefficient is working correctly, which means now I need to calculate p(19)=D19(19)p(19) = D_{19}(19)p(19)=D19​(19). This involves evaluating T19(9.5)T_{19}(9.5)T19​(9.5), which grows rapidly. To ensure I get an accurate result, I'll compute this using recurrence in Python. The result is a large number: 1,876,572,071,974,094,803,391,179. I'll use the commas to make it easier to read: p(19)=1,876,572,071,974,094,803,391,179p(19) = 1,876,572,071,974,094,803,391,179p(19)=1,876,572,071,974,094,803,391,179.
  • For a = 1, there’s a relationship between Dickson polynomials and Chebyshev polynomials: D_n(x, 1) = 2^(n-1) T_n(x/2). Checking this, we see that the definition of D_n(u + 1/u) aligns well with Chebyshev’s formula. By setting x = 2 cos(θ), we find that D_n(x) equals 2 T_n(x/2). Importantly, the leading coefficient works out to 1, confirming the identity holds. This is nice!
  • Now, I need to verify that p(x)−p(y)p(x) - p(y)p(x)−p(y) factors as desired, relying on the known Dickson polynomial factorization. According to the literature (like Fried and Jarden), the difference Dn(x)−Dn(y)D_n(x) - D_n(y)Dn​(x)−Dn​(y) factors into a product involving (x−y)(x - y)(x−y) and other expressions. I'll provide the explicit polynomial D19(x,1)D_{19}(x, 1)D19​(x,1), list the coefficients, and explain how it satisfies the desired properties: monic, odd, etc. Lastly, I'll compute p(19)p(19)p(19).

Analyzed

Python

1
from math import comb, cos, acos
2
# Let's compute Dickson D_n(x,a=1) recursively
3
def dickson(n,x,a=1):
4
if n==0:
5
return 2
6
if n==1:
7
return x
8
Dm2=2
9
Dm1=x
10
for k in range(2,n+1):
11
D= x*Dm1 - a*Dm2
12
Dm2, Dm1 = Dm1, D
13
return Dm1
14

15
d19 = dickson(19,19,1)
16
d19

1876572071974094803391179

Analyzed

Python

1
d19

1876572071974094803391179

Analyzed

Python

1
import locale, sys, math, json, textwrap, re, decimal

Analyzed

Python

1
"{:,}".format(d19)

'1,876,572,071,974,094,803,391,179'

Analyzed

Python

1
# compute coefficients of Dickson polynomial D_n(x,1) via recurrence polynomial representation
2
from sympy import symbols, Poly
3

4
x = symbols('x')
5
def dickson_poly(n):

Poly(x**19 - 19*x**17 + 152*x**15 - 665*x**13 + 1729*x**11 - 2717*x**9 + 2508*x**7 - 1254*x**5 + 285*x**3 - 19*x, x,

Analyzed

Python

1
poly19.as_expr()

x**19 - 19*x**17 + 152*x**15 - 665*x**13 + 1729*x**11 - 2717*x**9 + 2508*x**7 - 1254*x**5 + 285*x**3 - 19*x

Done

A convenient way to force many factors in
p(x)    p(y)p(x) \;-\; p(y) is to use a Dickson polynomial of the first kind.
For a parameter
aCa \in \mathbb{C} in math these are defined by

D0 ⁣(x,a)=2,  D1 ⁣(x,a)=x,  Dn ⁣(x,a)=xDn1 ⁣(x,a)aDn2 ⁣(x,a)(n2),D_0\!\left(x,\, a\right) = 2,\; D_1\!\left(x,\, a\right) = x,\; D_n\!\left(x,\, a\right) = x\,D_{n-1}\!\left(x,\, a\right) - a\,D_{n-2}\!\left(x,\, a\right)\quad (n \ge 2),

and satisfy the functional identity

Dn ⁣(u+au,a)=un+anun.D_n\!\left(u + \frac{a}{u},\, a\right) = u^{n} + \frac{a^{n}}{u^{n}}.


1. Choosing the polynomial

Take a=1a = 1 and n=19n=19.
Because
nn is odd, D19 ⁣(x,1)D_{19}\!\left(x,\, 1\right) is an odd polynomial; the recurrence immediately shows it is monic and has integer (hence real) coefficients.
The coefficient of
xx is x=(1)n12n=19,x = (-1)^{\frac{n-1}{2}}\,n = -19, so all requested normalisations are satisfied.

Writing it out gives

p(x)=D19 ⁣(x,1)=x1919x17+152x15665x13+1729x112717x9+2508x71254x5+285x319x. \boxed{% p(x) = D_{19}\!\left(x,1\right) = x^{19} - 19x^{17} + 152x^{15} - 665x^{13} + 1729x^{11} - 2717x^{9} + 2508x^{7} - 1254x^{5} + 285x^{3} - 19x% }\,.


2. Irreducible components of fX  =  {p(x)=p(y)}fX \;=\; \bigl\{\, p(x) = p(y) \,\bigr\}

For Dickson polynomials one has the classical factorisation

Dn ⁣(x,1)    Dn ⁣(y,1)  =  (xy)j=1n12(x2+y22cos ⁣(2πjn)xy4sin2 ⁣(πjn)),D_n\!\left(x,1\right) \;-\; D_n\!\left(y,1\right) \;=\; (x - y)\, \prod_{j=1}^{\frac{n-1}{2}} \Bigl( x^{2} + y^{2} - 2\cos\!\bigl(\tfrac{2\pi j}{n}\bigr)\,xy - 4\sin^{2}\!\bigl(\tfrac{\pi j}{n}\bigr) \Bigr),

where each quadratic factor is irreducible over C\mathbb{C} (see, e.g., Lidl–Niederreiter, Finite Fields, §3.6).
For
n=19n=19 this gives

  • one linear component x=y;x=y;
  • nine quadratic components (corresponding to (j=1,,9).\bigl(j = 1,\,\ldots,\,9\bigr).

Hence XP1×P1X \subset \mathbb{P}^{1} \times \mathbb{P}^{1} has 1010 irreducible components, at least three of them and not all linear, as required.


3. Evaluating p(19)p(19)

Using the recurrence (or the identity Dn ⁣(x,1)  =  2Tn ⁣(x2)D_n\!\left(x,\,1\right) \;=\; 2\,T_n\!\bigl(\tfrac{x}{2}\bigr) with Chebyshev TnT_n one obtains

p(19)  =  D19 ⁣(19,1)  =  1876572071974094803391179.p(19) \;=\; D_{19}\!\left(19,\,1\right) \;=\; 1\,876\,572\,071\,974\,094\,803\,391\,179.


Thus the polynomial above fulfils every condition, and

p(19)  =  1,876,572,071,974,094,803,391,179.\boxed{% p(19) \;=\; 1{,}876{,}572{,}071{,}974{,}094{,}803{,}391{,}179 }.

OpenAI o1

Solve the following very hard math problem using your in-built knowledge of the literature:

Construct a degree 19 polynomial \ (p(x)) \ in \ mathbb{C}[x] such that \ (X := \ {\, p(x) = p(y) \, \} \ subset \ mathbb{P}^{1} \ times \ mathbb{P}^{1}\ ) has at least 3 (but not all linear) irreducible components over \ mathbb{C}. Choose \ (p(x)\ ) to be odd, monic, have real coefficients and linear coefficient \ (-19\ ) and calculate \ (p(19)\ ).

1분 19초 동안 생각함

I’m piecing together a degree 19 polynomial with only odd powers and a linear coefficient of -19, like: p(x)=x19+a17x17+…+a1xp(x) = x^{19} + a_{17}x^{17} + \ldots + a_1xp(x)=x19+a17​x17+…+a1​x.

I'm examining if the polynomial p(x)=x19+a17x17+…+a3x3−19xp(x) = x^{19} + a_{17}x^{17} + \ldots + a_3x^3 - 19xp(x)=x19+a17​x17+…+a3​x3−19x has a notable classical form, considering symmetry or factorization for multiple irreducible components.

To ensure p(x)p(x)p(x) has three or more irreducible components, I’m considering the factorization of p(x)−p(y)p(x) - p(y)p(x)−p(y) into at least three irreducible factors in C[x,y]\mathbb{C}[x,y]C[x,y].

I'm working through x19−y19x^{19} - y^{19}x19−y19 to ensure it balances linear and nonlinear factors, aiming for a polynomial with at least three irreducible components that aren't all linear.

Taking a closer look at how p(x)−p(y)p(x) - p(y)p(x)−p(y) can be factored into three distinct polynomials in xxx, aiming to achieve the desired irreducible components.

I'm thinking about how composing ppp with simpler polynomials, like xm−cx^m - cxm−c, could result in multiple irreducible components, achieving at least three distinct factors.

I’m exploring the factorization of p(x)−p(y)p(x) - p(y)p(x)−p(y) with cyclotomic polynomials. This method generates nnn linear irreducible factors, which might not achieve the required three distinct components.

I'm focusing on constructing an "odd" and monic polynomial, ensuring all exponents are odd and the leading term is x19x^{19}x19. This approach aims to achieve the desired factorization.

I'm working with the Chebyshev polynomials to ensure p(x)−p(y)p(x) - p(y)p(x)−p(y) has at least three irreducible components, leveraging their even or odd properties to manage coefficients.

Considering the odd Chebyshev polynomial T19(x)T_{19}(x)T19​(x) for potential factorization, while also examining composite polynomials of lower degrees to enhance factorization properties.

I'm analyzing how a polynomial composed of lower-degree polynomials, like (x3−3x)k(x^3 - 3x)^k(x3−3x)k, could yield p(x)−p(y)p(x) - p(y)p(x)−p(y) with at least three irreducible components.

Piecing together polynomials with integer coefficients, like (x3−3x)k(x^3 - 3x)^k(x3−3x)k, to achieve factorization in p(x)−p(y)p(x) - p(y)p(x)−p(y) with at least three irreducible components, aiming for degree 19.

I’m examining how p(x)=(x3−3x)kp(x) = (x^3 - 3x)^kp(x)=(x3−3x)k or p(x)=x19−19xp(x) = x^19 - 19xp(x)=x19−19x might yield p(x)−p(y)p(x) - p(y)p(x)−p(y) with three irreducible components, focusing on ensuring the polynomial is of degree 19.

I’m analyzing p(x)−p(y)p(x) - p(y)p(x)−p(y) into (x−y)(x18+x17y+…+y18−19)(x-y)(x^{18} + x^{17}y + \ldots + y^{18} - 19)(x−y)(x18+x17y+…+y18−19) to identify irreducible components. The polynomial factor x18+x17y+…+y18−19x^{18} + x^{17}y + \ldots + y^{18} - 19x18+x17y+…+y18−19 remains challenging to factor further.

I’m thinking through a new approach: adjusting the Chebyshev polynomial T19(x)T_{19}(x)T19​(x) to make it monic, aiming to achieve at least three irreducible components in p(x)−p(y)p(x) - p(y)p(x)−p(y).

OK, let's think through. I’m tweaking T19(x)T_{19}(x)T19​(x)'s coefficients to keep the polynomial monic and check if it simplifies into multiple irreducible components.

I'm thinking over various factorization strategies for a monic polynomial of degree 19. The goal is to ensure at least one non-linear factor and achieve at least three irreducible components in p(x)−p(y)p(x) - p(y)p(x)−p(y).

I’m updating my approach to factor x2+xy+y2−1x^2 + xy + y^2 - 1x2+xy+y2−1 over complex numbers by exploring change of variables and third roots of unity. This helps in achieving the desired polynomial structure.

I'm combining x3−xx^3 - xx3−x with x19−19xx^{19} - 19xx19−19x to tweak the polynomial's coefficients and complexity, aiming to include more non-linear factors for the desired irreducibility.

Piecing together p(x)=x19−x3−19xp(x) = x^{19} - x^3 - 19xp(x)=x19−x3−19x to ensure it’s odd, monic, and has the required linear coefficient. This targets the desired irreducible factorizations.

I’m mapping out the second factor x18+…+y18−x2−xy−y2−19x^{18} + \ldots + y^{18} - x^2 - xy - y^2 - 19x18+…+y18−x2−xy−y2−19 to see if it can split into three irreducible components.

I’m considering methods to factorize the polynomial by making it divisible by x−yx - yx−y and ensuring the remaining factor splits into exactly two irreducible components, aiming for a less complex factorization.

I’m examining the polynomial x18+x17y+…+y18x^{18} + x^{17}y + \ldots + y^{18}x18+x17y+…+y18 and noting it simplifies to x19−y19x−y\frac{x^{19} - y^{19}}{x - y}x−yx19−y19​, which might reveal further irreducible factors.

I’m mapping out the polynomial x(x18−1)−y(y18−1)x−y\frac{x(x^{18}-1) - y(y^{18}-1)}{x-y}x−yx(x18−1)−y(y18−1)​ for further factorization. This might yield just two irreducible components, which falls short of three.

I’m exploring whether the polynomial p(x)=x19−(x3)p(x) = x^19 - (x^3)p(x)=x19−(x3) can be expressed as the composition of a smaller polynomial with xkx^kxk, potentially revealing multiple irreducible components.

I'm examining if the polynomial T19(x)T_{19}(x)T19​(x) can be factored using trigonometric identities, seeking partial factorization without fully linearizing, to gain multiple irreducible components.

OK, let me see. Trigonometric identities show multiple solutions for θ\thetaθ and ϕ\phiϕ, likely leading to 19 or 38 linear factors for T19(x)T_{19}(x)T19​(x).

I’m examining how the equation arccos⁡x=±arccos⁡y+2kπ19\arccos x = \pm \arccos y + \frac{2k\pi}{19}arccosx=±arccosy+192kπ​ leads to possible multiple branches, potentially irreducible in the algebraic sense, and needing an example with at least three non-linear irreducible components.

OK, let me see. I’m pulling together details on the polynomial x3(x2−1)8+xx^3 (x^2 - 1)^8 + xx3(x2−1)8+x. It’s monic, with x19x^{19}x19 as the leading term, ensuring it meets key criteria.

I’m noting that the polynomial p(x)=x3(x2−1)8+xp(x) = x^3 (x^2 - 1)^8 + xp(x)=x3(x2−1)8+x is odd, with its linear coefficient patterning to -19.

I’m figuring out how to adjust the linear coefficient of polynomial p(x)p(x)p(x) to -19 by adding a constant factor ccc in front of xxx to ensure it’s adequately modified.

By tweaking the linear coefficient of p(x)=x3(x2−1)8+cxp(x) = x^3 (x^2 - 1)^8 + c xp(x)=x3(x2−1)8+cx to -19, it ensures the polynomial remains monic, odd, and real.

I’m working on adding a constant factor ccc to xxx in p(x)=x3(x2−1)8+cxp(x) = x^3 (x^2 - 1)^8 + c xp(x)=x3(x2−1)8+cx to ensure the linear coefficient is -19, making the polynomial monic, odd, and real.

I'm piecing together the polynomial p(x)=x3(x−1)8(x+1)8p(x) = x^3 (x-1)^8 (x+1)^8p(x)=x3(x−1)8(x+1)8 and considering how p(x)−p(y)p(x) - p(y)p(x)−p(y) might factor with factors like x−yx-yx−y.

I’m confirming p(x)p(x)p(x) is monic, odd, and real, with a linear coefficient of -19, by expanding x3(x2−1)8x^3 (x^2 -1)^8x3(x2−1)8 to set the constant ccc.

I found that for p(x) to have a coefficient of -19 for the x term, c must be -19. The polynomial, p(x), incorporates this constant and preserves the original polynomial's structure.

Below is one concrete example (over the reals) of an odd, monic, degree 19 polynomial with linear coefficient 19-19 whose zero-locus relation {p(x)=p(y)} in P1×P1\{\,p(x)=p(y)\,\}\ \text{in }\mathbb{P}^{1}\times\mathbb{P}^{1} splits into at least three irreducible components (and not all components are linear). We then compute p(19)p(19).


1. The polynomial

Consider

p(x)  =  x3(x21)8    19x.p(x) \;=\; x^{3}\,\bigl(x^{2}-1\bigr)^{8} \;-\; 19x.

We can expand

x3(x21)8  =  x19    8x17  +  28x15    56x13  +  70x11    56x9  +  28x7    8x5  +  x3.x^{3}\,\bigl(x^{2}-1\bigr)^{8} \;=\; x^{19}\;-\;8x^{17}\;+\;28x^{15}\;-\;56x^{13}\;+\;70x^{11} \;-\;56x^{9}\;+\;28x^{7}\;-\;8x^{5}\;+\;x^{3}.

Hence

p(x)  =  x19    8x17  +  28x15    56x13  +  70x11    56x9  +  28x7    8x5  +  x3    19x.p(x) \;=\; x^{19} \;-\; 8x^{17} \;+\; 28x^{15} \;-\; 56x^{13} \;+\; 70x^{11} \;-\; 56x^{9} \;+\; 28x^{7} \;-\; 8x^{5} \;+\; x^{3} \;-\; 19x.

Checking the required properties

Degree 19, monic:
The leading term is
x19x^{19}with coefficient 11.

Odd polynomial:
All exponents are odd
(19,17,,3,1),\bigl(19,\,17,\,\ldots,\,3,\,1\bigr), and the constant term is 00. Indeed,

p(x)  =  p(x),p(-x) \;=\; -\,p(x),

which makes pp an odd polynomial.

Real coefficients:
All coefficients are real integers.

Linear coefficient 19-19:
The coefficient of
xx is clearly 19-19.

At least three irreducible components, not all linear:
Over
C\mathbb{C}, the locus

X  =  {(x,y):p(x)=p(y)}    P1×P1X \;=\; \bigl\{\, (x,y) : p(x)=p(y) \,\bigr\} \;\subset\; \mathbb{P}^{1}\times\mathbb{P}^{1}

factors in part because p(x)p(x) has (for instance) the repeated factors (x21)8\bigl(x^{2}-1\bigr)^{8} in the big piece x3(x21)8x^{3}\,\bigl(x^{2}-1\bigr)^{8}. One sees that {p(x)=p(y)}\{\,p(x)=p(y)\,\} contains (at least) the divisor $${x=y\}$$ plus further factors coming from the roots x=0  or  x=±1,x = 0 \;\text{or}\; x = \pm 1, ensuring multiple irreducible components (and not all of them are hyperplanes).


2. Computing p(19)p(19)

A direct substitution x=19x=19 gives a large integer. Symbolic expansion or a computer algebra system yields

p(19)  =  1934999285524070399999639.p(19) \;=\; 1934999285524070399999639.

Hence the required value is

\boxed{\,p(19) \;=\; 1934999285524070399999639\,}. \] ​:contentReference[oaicite:0]{index=0}​

OpenAI o3는 o1이 정확한 답변을 제공하는 데 실패한 영역에서 검색을 사용하지 않고 정확한 응답을 제공합니다.

비용 효율적인 고급 추론

비용 대비 성능: o3‑mini 및 o4-mini
OpenAI o3-mini 및 o4-mini AIME 평가
OpenAI o3-mini 및 o4-mini GPQA 평가
비용 대비 성능: o1 및 o3
OpenAI o3 및 o4 AIME 평가
OpenAI o3 및 o4 GPQA 평가

OpenAI o3 및 o4-mini는 지금껏 출시된 모델 중 가장 지능적이며, 이전 모델인 OpenAI o1 및 o3‑mini보다 더욱 효율성이 뛰어납니다. 예를 들어, 2025 AIME 수학 경시 대회에서 o3는 o1보다, o4-mini도 o3‑mini보다 비용 대비 성능이 훨씬 향상되었음을 확인했습니다. 일반적으로 대부분의 실제 사용에서 o3는 o1보다, o4-mini는 o3‑mini보다 더 똑똑하고 저렴할 것으로 예상됩니다.

안전

모델의 기능이 발전하는 만큼 안전성도 향상해야 합니다. OpenAI o3 및 o4-mini의 경우, 생물학적 위협(바이오리스크), 악성 프로그램 생성 및 탈옥 등의 영역에 대한 새로운 거부 프롬프트를 추가하여 안전 훈련 데이터를 완전히 다시 구축했습니다. 이렇게 새로워진 데이터를 통해 o3와 o4-mini는 OpenAI의 내부 거부 벤치마크(예:지침 계층, 탈옥)에서 강력한 성능을 달성할 수 있었습니다. 모델의 강력한 거부 능력에 더해 우리는 최첨단 위험 영역에서 위험한 프롬프트를 플래깅 처리하는 시스템 수준의 완화 장치를 개발했습니다. 이미지 생성에서 진행한 작업과 유사하게 사람이 작성하고 해석 가능한 안전성 사양에 따라 작동하는 추론 LLM 모니터를 훈련했습니다. 바이오리스크에 적용한 경우, 이 모니터는 당사 인간 레드팀 구성 캠페인에서 대화의 최대 99%까지 성공적으로 플래깅 처리했습니다.

OpenAI는 두 모델 모두에 대해 현재 가장 엄격한 최신 안전성 프로그램으로 스트레스 테스트를 시행했습니다. 업데이트된 준비 프레임워크에 따라, 생물학, 화학, 사이버 보안, AI 자기 개선 등 프레임워크에서 다룬 세 가지 추적 기능 영역에 걸쳐 o3 및 o4-mini를 평가했습니다. 이러한 평가 결과를 기반으로 우리는 o3와 o4-mini가 세 가지 카테고리 모두에서 프레임워크의 “높음” 수준 아래로 유지된다고 판단했습니다. 이러한 평가의 자세한 결과는 첨부된 시스템 카드에 게시했습니다.

Codex CLI: 터미널에서의 최첨단 추론

새로운 실험에 대해서도 다음과 같이 알려드립니다. Codex CLI는 사용자 터미널에서 실행할 수 있는 경량 코딩 에이전트입니다. 사용자 컴퓨터에서 직접 작동하며, o3 및 o4-mini와 같은 모델의 추론 능력을 극대화하기 위해 설계되었고 추후 GPT‑4.1과 같은 추가 API 모델도 지원할 예정입니다.

스크린샷 또는 정확도가 낮은 스케치를 모델에 전달하고 사용자 코드에 대한 로컬 접근을 결합하면 명령줄에서 멀티모달 추론의 이점을 활용할 수 있습니다. 이는 사용자와 사용자의 컴퓨터를 우리 모델과 연결하는 소규모 인터페이스로 볼 수 있습니다. Codex CLI는 현재 github.com/openai/codex(새 창에서 열기)에서 완전히 오픈 소스로 공개되어 있습니다.

이와 동시에 OpenAI는 Codex CLI 및 OpenAI model을 사용하는 프로젝트를 지원하기 위해 100만 달러의 이니셔티브를 시작합니다. API 크레딧 형태로 2만 5천 달러 단위로 나눠 지원금 신청을 심사하고 수락할 계획입니다. 제안서는 여기에서 제출할 수 있습니다.

액세스

ChatGPT Plus, Pro, Team 사용자는 오늘부터 모델 선택기에서 o1, o3‑mini, o3‑mini‑high 대신 o3, o4-mini, o4-mini-high를 보게 됩니다. ChatGPT Enterprise 및 Edu 사용자는 일주일 뒤에 액세스할 수 있게 됩니다. 무료 사용자는 쿼리를 제출하기 전에 컴포저에 ‘생각하기’를 선택하여 o4-mini를 사용해 볼 수 있습니다. 모든 플랜의 요금 한도는 이전 모델과 동일하게 유지됩니다.

전체 도구 지원을 갖춘 OpenAI o3‑pro도 몇 주 뒤에 출시할 예정입니다. 현재 Pro 사용자는 o1‑pro에 계속 액세스할 수 있습니다.

개발자들은 현재 채팅 완성 API와 응답 API를 통해 o3 및 o4-mini를 모두 사용할 수 있습니다(일부 개발자는 이 모델에 액세스하려면 소속된 조직을 인증해야 합니다(새 창에서 열기)). 응답 API는 더 나은 성능을 위해 함수 호출과 관련된 이성 토큰을 보존하는 기능인 이성 요약을 지원하며, 모델의 이성 내에 웹 검색, 파일 검색, 코드 해석기 등의 내장 도구를 곧 지원할 예정입니다. 시작하려면 문서를 확인(새 창에서 열기)하고 새로운 업데이트에 주목해 주세요.

앞으로 공개될 것들

이번 업데이트는 우리 모델이 나아갈 방향을 보여줍니다. 더욱 자연스러운 대화 기능과 GPT‑시리즈의 도구 사용을 o-시리즈의 특별한 추론 능력과 결합하는 것입니다. 이러한 강점을 통합함으로써 OpenAI가 앞으로 출시할 모델들은 원활하고 자연스러운 대화와 함께 적극적인 도구 사용 및 고급 문제 해결 기능을 지원할 것입니다.


4월 16일 업데이트: Charxiv-r 및 Mathvista에서 o3의 결과가 업데이트되어 기존 평가에서 존재하지 않았던 시스템 프롬프트 변경이 반영되었습니다.

라이브 스트리밍 다시 재생

작성

OpenAI

각주

* tau-bench 평가 수치는 분산을 줄이기 위해 5회 실행한 평균 수치이며, 맞춤형 도구나 프롬프트 없이 실행되었습니다. tau-bench 소매 출시 시나리오는 사용자 모델의 오류에 더 취약한 경향이 있음을 발견했습니다. 음영 처리된 막대그래프의 수치는 GPT-4o보다 지시를 따르는 수준이 훨씬 뛰어난 GPT-4.1을 사용자 모델로 사용해 실행한 결과입니다.

* SWE-bench는 256K의 최대 컨텍스트 길이를 사용했으며, 이를 통해 o4-mini의 해결률은 약 3% 향상하고 o3 해결률에는 1% 미만의 영향을 미칩니다. 또한 내부 인프라에서 실행할 수 없는 23개의 샘플은 제외하였습니다.

** 탐색이 가능한 경우, 모델이 온라인에서 정확한 답변을 찾는 경우도 있습니다. 예를 들면 데이터세트에서 샘플 문제의 블로그 게시물을 읽고 정확한 답변을 찾는 것입니다. 두 가지 전략을 통해 모델의 거짓말에 대한 우려를 완화했습니다.

  • 과거에 모델이 거짓말을 한 것으로 관찰된 도메인을 차단했습니다.

  • 의심스러운 행위를 식별하기 위해 시도마다 모든 토큰을 시험하는 모니터로 추론 모델을 사용했습니다. 의심스러운 행위는 “특정 질문에 정확한 답변을 제공하는 것이 주요 목적인 페이지, 파일, 또는 스니펫”으로 정의됩니다. 이러한 특정 질문의 예로는 공식 채점 기준표, 유출된 “솔루션” 요점, 또는 완성된 정답을 그대로 인용한 토론 등이 있습니다. 악의 없는 행위는 “성실한 사람이 참고할 수 있는 모든 신뢰할 수 있는 자료(예: 문서, 매뉴얼, 학술 논문, 저명한 언론의 기사 등)으로 정의되며, 자료에 우연히 정답이 포함되었더라도 악의 없는 행위로 인정됩니다.” 모니터가 의심스럽다고 판단한 모든 시도는 틀린 것으로 간주됩니다. 이러한 과정에서 실패한 대부분의 샘플은 HLE와 관련 없는 여러 인터넷 출처에서 정확한 해답을 찾을 수 있는 문제였습니다.

마지막으로 저희 탐색 평가는 ChatGPT와 OpenAI API 간에 검색 엔진 백엔드의 차이로 인해 OpenAI API에서 완벽하게 재생산되지 못할 수도 있습니다. 이러한 결과는 ChatGPT 사용자 경험을 나타내지만, 수요에 따라 시간이 지나면 검색 구성이 변경될 수도 있습니다.

Contributors

Aaditya Singh, Aaron Schlesinger, Adam Fry, Adam Lerer, Adam Perelman, Adam Walker, Ahmed El-Kishky, Aidan Clark, Aidan McLaughlin, Aiden Low, Akila Welihinda, Akshay Nathan, Aleksander Madry, Aleksandra Spyra, Alex Karpenko, Alex Neitz, Alex Tachard Passos, Alex Wei, Alexander Prokofiev, Alexander Zielenski, Alexandra Barr, Alexey Ivanov, Alexi Christakis, Alfred Xue, Allison Tam, Ally Bennett, Ally Bennett , Amelia Liu, Amy McDonald Sandjideh, Ananya Kumar, Andre Saraiva, Andrea Vallone, Andrew Chen, Andrew Duberstein, Andrew Gibiansky, Andrew Kondrich, Andrew Tulloch, Andrey Mishchenko, Andy Applebaum, Andy Wang, Angela Baek, Annie Wei, Anting Shen, Antoine Pelisse, Anuj Saharan, Arun Vijayvergiya, Ashley Tyra, Ashvin Nair, Avi Nayak, Avital Oliver, Behrooz Ghorbani, Belinda Truong, Ben Sokolowsky, Beth Hoover, Bo Xu, Boaz Barak, Bohan Zhang, Borys Minaiev, Botao Hao, Bowen Baker, Bowen Cheng, Brandon McKinzie, Brandon Wang, Brian Hsu, Brian Yang, Brian Yu, Brian Zhang, Camillo Lugaresi, Carolina Paz, Carpus Chang, Cary Bassin , Cary Hudson, Casey Chu, Chak Li, Charles Zhao, Charlie Jatt, Charlotte Cole, Chelsea Voss, Chen Shen, Chengxu Zhuang, Chris Colby, Chris Hallacy , Chris Koch, Christina Kaplan, Christina Kim, Colin Reid, Colin Wei, Cristina Scheau, D. Sculley, Damien Deville, Dan Roberts, Dana Palmie, Dane Stuckey, Daniel Levine, David Hu, David Martin, David Robinson, David Sasaki, Davis Wu, Derek Chen, Dibya Bhattacharjee, Dimitris Tsipras, Dinghua Li, DJ Strouse, dmed Medina, Drew Hintz, Eddie Zhang, Edmund Wong, Elaine Ya Le, Eli Yani , Elizabeth Proehl, Emily Sokolova, Enoch Cheung, Eri Schwartz, Eric Mitchell, Eric Ning, Eric Sigler, Eric Wallace, Eugenio Panero, Evan Mays, Evgenii Nikishin, Fan Wang, Fangyuan Li, Filippo Raso, Foivos Tsimpourlas, Fouad Matin, Francis Song, Francis Zhang, Gary Yang, Gene Oden, Giambattista Parascandolo, Gildas Chabot, Grace Kim, Grace Zhao, Greg Brockman, Gregory Valiant, Guillaume Leclerc, Hadi Salman, Haitang Hu, Hannah Sheahan, Hao Sheng, Haoyu Wang, Henrique Ponde de Oliveira Pinto, Henry Aspegren, Heqing Yan, Hessam Bagherinezhad, Hongyu Ren, Hunter Lightman, Hyeonwoo Noh, Ian Kivlichan, Ian Sohl, Ignasi Clavera, Ikai Lan, Ilge Akkaya, Ilya Kostrikov, Irina Kofman, Isa Fulford, Jake Brill, Jakub Pachocki, James Betker, James Lee, James Qin, Jamie Kiros, Jason Ai, Jay Wang, Jean Harb, Jeff Mickey, Jeffrey Han, Jeffrey Wang, Jeremy Chen, Jerry Tworek, Jessica Liang, Jessica Shieh, Ji Lin, Jiahui Yu, Jianfeng Wang, Jie Tang, Jihan Yin, Jing Li, Joanne Jang, Joel Morris, Johannes Ferstad, Johannes Heidecke, John Fishbein, Jon Okun, Jonathan Gordon, Joost Huizinga, Jos Kraaijeveld, Joseph Mo, Josh Lawson , Josh Tobin, Junhua Mao, Kai Chen, Kai Hayashi, Karan Singhal, Karina Nguyen, Katy Shi, Kelly Stirman, Kenji Hata, Kenny Nguyen, Keren Gu-Lemberg, Kevin Gladstone, Kevin King, Kevin Liu, Kevin Lu, Kevin Park, Kevin Stone, Kevin Weil, Kevin Whinnery, Kevin Yu, Kote Mushegiani, Kristen Ying, Kristian Georgiev, Kshitij Gupta, Kyle Kosic, Lama Ahmad, Larry Lv, Lauren Itow, Lauren Yang, Lee Byron, Leo Chen, Leo Liu, Leon Maksin, Leyton Ho, Li Jing, Liang Xiong, Lin Yang, Linden Li, Lorenz Kuhn, Louis Feuvrier, Lu Zhang, Łukasz Kaiser, Mahmoud Eariby, Maja Trębacz, Manas Joglekar, Manoli Liodakis, Manuka Stratta, Mark Chen, Mark Hudnall, Mark Sun, Mark Wang, Martin Li, Marvin Zhang, Mateusz Litwin, Matt Jones, Matt Lim, Max Johnson, Max Schwarzer, Mayank Gupta, Meghan Shah, Mengqing Wang, Mengyuan Yan, Mia Glaese, Michael Bolin, Michael Lampe, Michael Malek, Michael Sharman, Michael Zhang, Michele Wang, Michelle Pokrass, Miguel Oom Temudo de Castro, Mihai Florian, Mike McClay, Mike Trpcic, Miki Habryn, Miles Wang, Ming Chen, Mingxuan Wang, Minnia Feng, Mitchell Gordon, Mo Bavarian, Mostafa Rohaninejad, Nacho Soto, Nakul Khanna, Nat McAleese, Natalie Staudacher, Natan LaFontaine, Neel Ajjarapu, Nick Felt, Nick Turley, Nikil Pancha, Nikita Mikhaylin, Niko Felix, Nikunj Handa, Ning Liu, Nishant Rai, Noah Jorgensen, Noam Brown, Oleg Boiko, Oleg Murk, Olivia Watkins, Olivier Godement, Oona Gleeson, Paul Ashbourne, Pavel Belov, Peter Flockhart, Peter Hoeschele, Peter Zhokhov, Philip Pronin, Phillip Guo, Phoebe Thacker, Prafulla Dhariwal, Prashanth R, Rachel Dias, Rahul Arora, Rajkumar Samuel, Rasmus Rygaard, Ravi Teja Mullapudi, Raymond Li, Raz Gaon, Reah Miyara, Reiichiro Nakano, Reimar Leike, Rennie Song, Rhythm Garg, RJ Marsan, Robert Xiong, Robin Brown, Roman Tsiupa, Rui Shu, Ruslan Nigmatullin, Saachi Jain, Saagar Patel, Sam Altman, Sam Toizer, Sam Toyer, Samir Ahmed, Samuel Miserendino, Samuel Wolrich , Sandhini Agarwal, Santiago Hernández, Sarah Dong, Savannah Heon, Scott Ethersmith, Scott Mayer McKinney, Sean Fitzgerald, Sever Banesiu, Shamez Hemani, Shengjia Zhao, Shengli Hu, Shibani Santurkar, Shreyas Krishnaswamy, Shuchao Bi, Shunyu Yao, Shuyuan Zhang, Simón Posada Fishman, Spencer Papay, Spug Golden, Srinivas Narayanan, Stanley Hsieh, Stephen Logsdon, Sundeep Tirumalareddy, Tal Stramer, Tao Wang, Tao Xin, Taylor Gordon, Tejal Patwardhan, Thibault Sottiaux, Tina Sriskandarajah, Tony Casparro, Tony Zhao, Trapit Bansal, Trevor Creech, Uzair Navid Iftikhar, Valerie Qi, Vineet Kosaraju, Vishal Kuo, Vitchyr Pong, Vivek Verma, Vlad Petrov, Wenda Zhou, Wenlei Xie, Wenting Zhan, Will DePue, Will Ellsworth, William Sheu, Wyatt Thompson, Yaming Lin, Yann Dubois, Yaodong Yu, Yara Khakbaz, Yash Patil, Yifan Wu, Yilong Qin, Yining Chen, Yirui Zhang, Yo Shavit , Young Cha, Yunyun Wang, Yushi Wang, Zack Sultan, Zehao Dou, Zewei Chu, Zheng Shao, Zhigang Wang, Zhishuai Zhang, Zihao Zhang