ወደ ዋና ይዘት እለፍ
OpenAI
በመጫን ላይ…

እ.ኤ.አ ጁን 10፣ 2025 የዘመነ፦ OpenAI o3‑pro አሁን በChatGPT ውስጥ እንዲሁም በእኛ API ውስጥ ለPro ተጠቃሚዎች ይገኛልልክ እንደ OpenAI o1‑pro፣ o3‑pro እጅግ የማሰብ ችሎታ ያለው የሞዴላችን ስሪት ማለትም OpenAI o3 ሲሆን ይህም ረዘም ላለ ጊዜ እንዲያስብ እና አስተማማኝ ምላሽ መስጠት እንዲችል ታስቦ የተሠራ ነው። ሙሉ ዝርዝሮች በእኛ የሚለቀቁ ማስታወሻዎች (በአዲስ መስኮት ውስጥ ይክፈታል)ውስጥ ሊገኙ ይችላሉ


ዛሬ ምላሽ ከመስጠት በፊት ረዘም ላለ ጊዜ እንዲያስቡ የሰለጠኑ በእኛ የo-ተከታታይ ሞዴሎች ውስጥ የቅርብ ጊዜዎቹን OpenAI o3 እና o4-miniን እየለቀቅን ነው። እነዚህ እስከዛሬ ከተለቀቁት ሞዴሎች እጅግ ብልህ የሆኑ ሞዴሎች ናቸው፣ ይህም ጉጉት ካላቸው ተጠቃሚዎች እስከ የላቁ ተመራማሪዎች ድረስ ለሁሉም ሰው የChatGPTን አቅሞች የሚቀይር ትልቅ እርምጃን ያመለክታሉ። ለመጀመሪያ ጊዜ፣ የማመዛዘን ሞዴሎቻችን በChatGPT ውስጥ የሚገኝ ማንኛውንም መሣሪያ በወኪል መጠቀም እና ማዋሃድ ይችላሉ—ይህ ድርን መፈለግ፣ የተሰቀሉ ፋይሎችን እና ሌሎች መረጃዎችን በPython መተንተን፣ ስለ እይታ ግቤቶች በጥልቀት ማመዛዘን እና ምስሎችን ማመንጨትን እንኳን ያካትታል። በወሳኝ ሁኔታ፣ እነዚህ ሞዴሎች በተለምዶ ከአንድ ደቂቃ ባነሰ ጊዜ ውስጥ ይበልጥ ውስብስብ የሆኑ ችግሮችን ለመፍታት፣ በትክክለኛ የውጤት ቅርጸቶች ዝርዝር እና የታሰቡ መልሶችን ለማምረት መሣሪያዎችን መቼ እና እንዴት መጠቀም እንዳለባቸው ለማመዛዘን የሰለጠኑ ናቸው። ይህ ብዙ ገጽታ ያላቸውን ጥያቄዎች በበለጠ ብቃት እንዲፈቱ ያስችላቸዋል፣ ይህም እርስዎን ወክሎ በተናጥል ተግባራትን ወደሚያከናውን ወደ ይበልጥ ወኪል የሆነ ChatGPT የሚወስድ እርምጃ ነው። የተራቀቀ የማመዛዘን ኃይል ከሙሉ የመሣሪያ መዳረሻ ጋር ተጣምሮ በትምህርታዊ መመዘኛዎች እና በእውነተኛ ዓለም ተግባራት ላይ በጣም ጠንካራ አፈጻጸም ያስገኛል፣ ይህም በሁለቱም በማሰብ ችሎታ እና በጠቃሚነት ላይ አዲስ ደረጃን ያስቀምጣል።

ምን ተለወጠ

OpenAI o3 ድንበርን በኮዲንግ፣ በሂሳብ፣ በሳይንስ፣ በእይታ ግንዛቤ እና በሌሎችም ላይ የሚገፋፋ የእኛ በጣም ኃይለኛ የማመዛዘን ሞዴል ነው። Codeforces፣ SWE-bench (ብጁ ሞዴል-ተኮር ስካፎል ሳይገነባ) እና MMMUን ጨምሮ አዲስ SOTA በማመሳከሪያዎች ላይ ያዘጋጃል። ባለብዙ ገፅታ ትንተና ለሚፈልጉ እና ምላሾቹ ወዲያውኑ ግልጽ ላይሆኑ ለሚችሉ ውስብስብ ጥያቄዎች ተስማሚ ነው። በተለይም ምስሎችን፣ ገበታዎችን እና ግራፊክስን በመተንተን በእይታ ስራዎች ላይ በብርቱ ይሰራል። በውጪ ባለሙያዎች ግምገማ ውስጥ፣ o3 በከባድ እና በአስቸጋሪ ዓለም ተግባራት ላይ ከOpenAI o1 በ20 በመቶ ያነሰ ዋነኛ ስህተቶችን ያደርጋል፤ በተለይም በፕሮግራሚንግ፣ ንግድ ሥራ/ማማከር እና የፈጠራ አስተሳሰብ ዘርፎች ውስጥ የላቀ ነው። ቀደምት ሞካሪዎች የትንታኔ ጥንካሬውን እንደ የአስተሳሰብ አጋር አጉልተው አሳይተዋል እና አዲስ ሀሳቦችን ያመነጫሉ እና በትክክል ማገምገም ችሎታውን አጽንዖት ሰጥተዋል—በተለይ በባዮሎጂ፣ ሂሳብ እና ምህንድስና አውዶች።

OpenAI o4-mini ለፈጣን እና ወጪ ቆጣቢ ምክኒያት የተመቻቸ አነስተኛ ሞዴል ነው—በመጠን እና በዋጋ በተለይም በሂሳብ፣ በኮዲንግ እና በምስል ተግባሮች ላይ አስደናቂ አፈፃፀም ያለው። በAIME 2024 እና 2025 ላይ ምርጥ አፈጻጸም ያለው የቤንችማርክ ሞዴል ነው። ምንም እንኳን የኮምፒዩተር መዳረሻ የAIME ፈተናን አስቸጋሪነት በትክክል ቢቀንስም፣ እኛም እንደገና አስተዋልን እንደ o4-mini በAIME 2025 ላይ 99.5% pass@1 (100% consensus@8) ማሳካቱ በፓይዘን አስተርጓሚ ሲሰጥ ተገኝቷል። እነዚህ ውጤቶች ከመሳሪያ መዳረሻ ያላቸው ሞዴሎች አፈጻጸም ጋር መወዳደር አይገባም፣ ነገር ግን o4-mini ያሉትን መሳሪያዎች በተገቢ ሁኔታ እንደሚጠቀም የሚያሳይ አንድ ምሳሌ ናቸው፤ o3 በAIME 2025 ላይ ተመሳሳይ ማሻሻያዎችን ከመሳሪያ አጠቃቀም ያሳያል (98.4% pass@1፣ 100% consensus@8)።

በባለሙያ ግምገማዎች ውስጥ፣ o4-mini እንደ ዳታ ሳይንስ ያሉ ጎራዎች እና ከSTEM ውጪ በሆኑ ተግባራት ላይ ከቀዳሚው o3‑mini ይበልጣል። ለውጤታማነቱ ምስጋና ይግባውና o4-mini ከo3 በከፍተኛ ደረጃ የአጠቃቀም ገደቦችን ድጋፍ ያደርጋል፣ ይህም ከማመዛዘንነት ለሚጠቅሙ ጥያቄዎች ጠንካራ ከፍተኛ የማስተላለፊያ ዘዴ አማራጭ ነው። የውጪ ኤክስፐርት ገምጋሚዎች ለተሻሻለ ብልህነት እና የድር ምንጮችን በማካተት ሁለቱንም ሞዴሎች የተሻሻለ መመሪያን በማሳየት እና ከቀደምቶቹ የበለጠ ጠቃሚ እና ሊረጋገጡ የሚችሉ ምላሾችን አሳይተዋል። ከቀደምት የማመዛዘን ሞዴሎቻችን ጋር ሲነጻጸር፣ እነዚህ ሁለቱ ሞዴሎችም የበለጠ ተፈጥሯዊ እና ውይይቶች ሊሰማቸው ይገባል፣ በተለይም የማስታወስ እና ያለፉ ንግግሮችን በማጣቀስ ምላሾችን የበለጠ ግላዊ እና ተዛማጅነት ያለው ለማድረግ።

አይነተ ብዙ

ኮዲንግ

ሁሉም የSWE-bench ግምገማ ስራዎች በውስጣዊ መሠረተ ልማት ላይ እንደተረጋገጡ ተግባራት n=477 ተግባሮችን ቋሚ ንዑስ ስብስብ ይጠቀማሉ።

መመሪያ መከተል እና የወኪል መሳሪያዎች አጠቃቀም

ሁሉም ሞዴሎች በከፍተኛ ‹የማመዛዘን ጥረት› ቅንጅቶች ይገመገማሉ—በChatGPT ውስጥ ካሉ ‹o4-mini-high› ልዩነቶች ጋር ተመሳሳይ ነው።

የማጠናከሪያ ትምህርትን ማስፋፋት መቀጠል

በOpenAI o3 እድገት ወቅት፣ መጠነ ሰፊው የማጠናከሪያ ትምህርት በGPT‑ተከታታይ ቅድመ ስልጠና ውስጥ የሚታየውን ተመሳሳይ «ተጨማሪ ስሌት = የተሻለ አፈጻጸም» እንደሚያሳይ አስተውለናል። የመለኪያ መንገዱን በመከታተል—በዚህ ጊዜ በRL ውስጥ—በሁለቱም በስልጠና ስሌት እና በማጣቀሻ-ጊዜ ማመዛዘን ውስጥ ተጨማሪ መጠን ወዳለው ትዕዛዝ ገፍተናል፣ ነገር ግን አሁንም ግልጽ የሆኑ የአፈጻጸም ብልጫዎችን እያየን ነው፣ ይህም ሞዴሎቹ የበለጠ እንዲያስቡ በተፈቀደላቸው መጠን የአፈጻጸማቸው ሁኔታ እየተሻሻለ መሄዱን ያረጋግጣል። ከOpenAI o1 ጋር እኩል በሆነ መዘግየት እና ወጪ፣ o3 በChatGPT ውስጥ ከፍተኛ አፈጻጸም ይሰጣል—እና ረዘም ላለ ጊዜ እንዲያስብ ከፈቀድንለት አፈጻጸሙ እየጨመረ እንደሚሄድ አረጋግጠናል።

በተጨማሪም ሁለቱንም ሞዴሎች በማጠናከሪያ ትምህርት አማካኝነት መሣሪያዎችን እንዲጠቀሙ አሠልጥነናል—መሣሪያዎችን እንዴት እንደሚጠቀሙ ብቻ ሳይሆን መቼ መጠቀም እንዳለባቸው እንዲያመዛዝኑም አስተምረናል። በተፈለገው ውጤት ላይ በመመርኮዝ መሣሪያዎችን የማሰማራት ችሎታቸው ክፍት በሆኑ ሁኔታዎች ውስጥ የበለጠ ችሎታ እንዲኖራቸው ያደርጋቸዋል—ይህም በተለይም የእይታ ማመዛዘን እና ባለብዙ እርምጃ የሥራ ፍሰቶችን የሚያካትቱ ሁኔታዎች ውስጥ ነው። በቀደምት ፈታኞች ሪፖርት በተደረገው መሰረት፣ ይህ መሻሻል በሁለቱም በትምህርታዊ መመዘኛዎች እና በእውነተኛ ዓለም ተግባራት ላይ ይንጸባረቃል።

ምስሎችን በመጠቀም በማሰብ ላይ

ChatGPT thinking with images

ለመጀመሪያ ጊዜ እነዚህ ሞዴሎች ምስሎችን በቀጥታ ወደ የሃሳብ ሰንሰለታቸው ማዋሃድ ይችላሉ። ምሥልን ብቻ አያዩም—አብረው ያስባሉ። ይህ በመልቲሞዳል መመዘኛዎች ላይ በዘመናዊ አፈፃፀማቸው ላይ የሚንፀባረቅ ምስላዊ እና ጽሑፋዊ አስተሳሰብን የሚያዋህድ አዲስ የችግር ፈቺ ክፍልን ይከፍታል።

ሰዎች የነጭ ሰሌዳ፣ የመማሪያ መጽሃፍ ዲያግራም ወይም በእጅ የተሳለ ንድፍ ፎቶ ሊስቀሉ ይችላሉ፣ እና ሞዴሉ ምስሉ ብዥታ፣ የተገለበጠ ወይም ዝቅተኛ ጥራት ቢሆንም—ሊተረጉመው ይችላል። በመሣሪያ አጠቃቀም፣ ሞዴሎች ምስሎችን በቀላሉ መቆጣጠር ይችላሉ—ማሽከርከር፣ ማጉላት ወይም መቀየር እንደ ማመዛዘን ሂደታቸው መለወጥ ይችላሉ።

እነዚህ ሞዴሎች በእይታ ግንዛቤ ተግባራት ላይ ምርጡን ትክክለኛነትን ይሰጣሉ፣ ይህም ከዚህ በፊት ሊደረሱባቸው የማይችሉ ጥያቄዎችን ለመፍታት ያስችላሉ። ተጨማሪ ለማወቅ የእይታ ማመዛዘን ምርምር ብሎግ የሚለውን ይመልከቱ።

ወደ ወኪል መሣሪያ አጠቃቀም

OpenAI o3 እና o4-mini በChatGPT ውስጥ ያሉ መሳሪያዎችን እና በAPI ውስጥ በተግባራዊ ጥሪ የእራስዎን ብጁ መሳሪያዎች ሙሉ በሙሉ ማግኘት ይችላሉ። እነዚህ ሞዴሎች ችግሮችን እንዴት መፍታት እንደሚችሉ፣ ዝርዝር እና አሳቢ መልሶችን በትክክለኛው የውጤት ቅርጸቶች በፍጥነት ለማምረት መቼ እና እንዴት እንደሚመርጡ በመምረጥ የሰለጠኑ ናቸው—በተለምዶ ከአንድ ደቂቃ በታች።

ለምሳሌ፣ አንድ ተጠቃሚ እንዲህ ሊጠይቅ ይችላል፦ «በCalifornia የበጋ የኢነርጂ አጠቃቀም ከባለፈው ዓመት ጋር እንዴት ይነጻጸራል?» ሞዴሉ ድሩን ለህዝብ አገልግሎት መረጃ መፈለግ፣ ትንበያ ለመገንባት የፓይተን ኮድ መጻፍ፣ ግራፍ ወይም ምስል ያመነጫል እና ከትንበያው በስተጀርባ ያሉትን ቁልፍ ነገሮች ማብራራት ይችላል፣ ብዙ የመሳሪያ ጥሪዎችን አንድ ላይ በማያያዝ። ማመዛዘን ሞዴሎች ለሚገናኙት መረጃ መሠረት እንደ አስፈላጊነቱ ምላሽ ማቅረብ እና መቀየር ይችላሉ። ለምሳሌ፣ በፍለጋ አቅራቢዎች እርዳታ ድሩን ብዙ ጊዜ መፈለግ፣ ውጤቶችን መመልከት እና ተጨማሪ መረጃ ከፈለጉ አዲስ ፍለጋዎችን መሞከር ይችላሉ።

ይህ ተለዋዋጭ እና ስልታዊ አቀራረብ ሞዴሎችን ከሞዴሉ ውስጥ ያለውን እውቀት በማሻገር ወቅታዊ መረጃ ማግኘት፣ የተራዘመ ማመዛዘንግ፣ ውህደት እና ውጤት ማመንጨት የሚጠይቁ ተግባራትን እንዲፈቱ ያስችላቸዋል።

ሁሉም ምሳሌዎች በOpenAI o3 ተጠናቀዋል።

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)\ ).

በምክንያት የተደገፈ ስለ ፖሊኖሚል ግንባታ ለ55 ሰከንድ
  • 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):

ፖሊ(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)\ ).

ሐሳብ ለ1 ደቂቃ 19 ሰከንድ

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 o3 ፍለጋ ሳይጠቀም ምላሹን በትክክል ያገኛል፣ ነገር ግን o1 ትክክለኛ ምላሽ መስጠት አልቻለም።

ወጪ ቆጣቢ አስተሳሰብን ማመዛዘን

ወጪ እና አፈጻጸም: o3‑mini እና o4-mini
OpenAI o3-mini እና o4-mini AIME ኢቫልስ
OpenAI o3-mini እና o4-mini GPQA ኢቫልስ
ወጪ እና አፈጻጸም፦ o1 እና o3
OpenAI o3 እና o4 AIME ኢቫልስ
OpenAI o3 እና o4 GPQA ኢቫልስ

OpenAI o3 እና o4-mini እስካሁን የለቀቅነው በጣም አስተዋይ ሞዴሎች ናቸው፣ እና እንዲሁም ከቀደምቶቹ OpenAI o1 እና o3‑mini የበለጠ ቀልጣፋ ናቸው። ለምሳሌ፣ በ2025 AIME የሂሳብ ውድድር፣ የOpenAI o3 ወጪ-አፈጻጸም ድንበር ከOpenAI o1 በላይ በጥብቅ ይሻሻላል፣ እና በተመሳሳይ መልኩ፣ የOpenAI o4-mini ድንበር ከOpenAI o3‑mini በጥብቅ ይሻሻላል። በአጠቃላይ፣ ለአብዛኛው የገሃዱ ዓለም አጠቃቀም፣ o3 እና o4-mini እንዲሁ ከo1 እና o3‑mini የበለጠ ብልህ እና ርካሽ ይሆናሉ ብለን እንጠብቃለን።

ደህንነት

በሞዴል አቅሞች ውስጥ የሚደረግ እያንዳንዱ ማሻሻያ ለደኅንነት ተመጣጣኝ ማሻሻያዎችን ያረጋግጣል። ለOpenAI o3 እና o4-mini፣ የእኛን የደኅንነት የስልጠና ውሂብ ሙሉ በሙሉ እንደገና ገንብተናል፣ ይህም እንደ ባዮሎጂያዊ ስጋቶች (ባዮስጋት)፣ ተንኮል አዘል ሶፍትዌር ማመንጨት እና የሶፍትዌር ገደቦችን ለተንኮል ዓላማ መተላለፍ (ጄልብሬኮች) ባሉ አካባቢዎች አዲስ የመቃወም እርምጃዎችን አክለናል። ይህ የታደሰው ውሂብ o3 እና o4-mini ሞዴሎች በውስጣዊ የመቃወም መመዘኛዎቻችን ላይ (ለምሳሌ፣ የመመሪያ ተዋረድ፣ የሶፍትዌር ገደቦችን ለተንኮል ዓላማ መተላለፍ (ጄልብሬኮች)) ጠንካራ አፈጻጸም እንዲያሳኩ አድርጓቸዋል። ለሞዴል መቃወሞች ጠንካራ አፈጻጸም ከማሳየት በተጨማሪ፣ በድንበር ላይ ባሉ የስጋት አካባቢዎች አደገኛ እርምጃዎች ላይ ምልክት ለማድረግ በሥርዓት ደረጃ ቅነሳዎችን አዘጋጅተናል። ቀደም ሲል በምስል ማመንጨት ውስጥ እንደሠራነው ሁሉ በተመሳሳይ መልኩ፣ ከሰው በተጻፉ እና ሊተረጎሙ ከሚችሉ የደኅንነት መስፈርቶች የሚሠራ የማመዛዘን LLM መቆጣጠሪያ አሰልጥነናል። በባዮስጋት ላይ ሲተገበር፣ ይህ መቆጣጠሪያ በሰው ቀይ‑የቡድን ዘመቻችን ውስጥ ~99% የሚሆኑ ውይይቶችን በተሳካ ሁኔታ ምልክት አድርጓል።

እስከ አሁን ድረስ ሁለቱንም ሞዴሎች በጣም ጥብቅ በሆነው የደኅንነት ፕሮግራማችን በጥልቀት ፈትነናቸዋል። በዘመነው የዝግጁነት ማዕቀፍ መሰረት፣ o3 እና o4-ሚኒን በማዕቀፉ በተካተቱ ሦስት የአቅም ዝርፎች ማለትም፦ ባዮሎጂካል እና ኬሚካል፣ የሳይበር ደኅንነት እና የሰው ሠራሽ አስተውሎት (AI) ራስን መሻሻል ላይ ባላቸው አፈጻጸም ገምግመናል። በእነዚህ የግምገማ ውጤቶች መሠረት፣ ሁለቱም o3 እና o4‑mini በሦስቱም ምድቦች ውስጥ ከማዕቀፉ «ከፍተኛ» ደረጃ በታች እንደሚቆዩ ወስነናል። የእነዚህን ግምገማዎች ዝርዝር ውጤቶችን በተያያዡ የሥርዓት ካርድ⁠ ውስጥ አትመናል።

Codex CLI፦ በተርሚናል ውስጥ የድንበር ማመዛዘን

እንዲሁም አዲስ ሙከራን እያጋራን ነን፦ Codex CLI፣ ቀላል ክብደት ያለው የኮድ ወኪል ከእርስዎ ተርሚናል ላይ ማስኬድ ይችላሉ። በቀጥታ በኮምፒዩተርዎ ላይ ይሰራል እና እንደ o3 እና o4-mini ያሉ ሞዴሎች የማስተዋል አቅሞችን ለማስበረቅ የተነደፈ ሲሆን በቅርብ ጊዜ ለተጨማሪ የAPI ሞዴሎች ድጋፍ እንደ GPT‑4.1 ይሰጣል።

ቅጽበታዊ ገጽ እይታዎችን ወይም ዝቅተኛ ታማኝነት ንድፎችን ወደ ሞዴሉ በማለፍ የመልቲሞዳል ማመዛዘንን ጥቅሞችን ከትእዛዝ መስመሩ ማግኘት ይችላሉ። እኛ ሞዴሎቻችንን ከተጠቃሚዎች እና ከኮምፒውተሮቻቸው ጋር ለመአገናኝ እንደ አነስተኛ በይነገጽ እናስባለን። Codex CLI በgithub.com/openai/codex(በአዲስ መስኮት ውስጥ ይክፈታል) ላይ በሙሉ ክፍት-ምንጭ ነው ዛሬ።

ከዚህ ጎን ለጎን Codex CLI እና OpenAI ሞዴሎችን በመጠቀም ፕሮጀክቶችን ድጋፍ የ$1 ሚሊዮን ተነሳሽነት እንጀምራለን። በAPI ክሬዲቶች መልክ በ$25,000 ጭማሪ የሚገኙ መተግበሪያዎችን እንገመግማለን እና እንቀበላለን። ሀሳቦችን እዚህ ሊያስረክቡ ይችላሉ።

መዳረሻ

የChatGPT Plus፣ Pro እና የቡድን ተጠቃሚዎች ከዛሬ ጀምሮ በሞዴል መራጭ ውስጥ o3፣ o4-mini እና o4-mini-high ያያሉ፣ ይህም o1፣ o3‑mini እና o3‑mini‑highን ይተካሉ። ChatGPT Enterprise እና የEdu ተጠቃሚዎች በአንድ ሳምንት ውስጥ መዳረሻ ያገኛሉ። ነፃ ተጠቃሚዎች ጥያቄያቸውን ከማቅረባቸው በፊት በአቀናባሪው ውስጥ ‹Think› የሚለውን በመምረጥ o4-mini መሞከር ይችላሉ። በሁሉም እቅዶች ላይ ያሉ የመጠን ገደቦች ከቀደምት የሞዴሎች ስብስብ ሳይለወጡ ይቆያሉ።

በጥቂት ሳምንታት ውስጥ OpenAI o3‑proን ከሙሉ የመሳሪያ ድጋፍ ጋር ለመልቀቅ እንጠብቃለን። ለአሁን፣ የPro ተጠቃሚዎች አሁንም o1‑pro መዳረሻ አላቸው።

ሁለቱም o3 እና o4-mini ዛሬ በውይይት Completions API እና Responses API በኩል ለአበልፃጊዎች ይገኛሉ (አንዳንድ አበልፃጊዎች እነዚህን ሞዴሎች ለመድረስ ድርጅቶቻቸውን አረጋግጥ(በአዲስ መስኮት ውስጥ ይክፈታል) አለባቸው)። የምላሾች API ማጠቃለያዎችን ይደጋፋል፣ በተግባር ዙሪያ የማሰብ token-ዎችን ለተሻለ አፈጻጸም ይጠብቃል፣ እና በቅርቡ እንደ ድር ፍለጋ፣ ፋይል ፍለጋ እና ኮድ ተርጓሚ በሞዴል ማሰብ ውስጥ የተገነባ መሳሪያዎችን ይደጋፋል። ለመጀመር ሰነዶቻችንን ማሰስ(በአዲስ መስኮት ውስጥ ይክፈታል) እና ለተጨማሪ ዝመናዎች ይከታተሉ።

ቀጣይ ምን ይሆናል?

የዛሬዎቹ ዝማኔዎች ሞዴሎቻችን እየሄዱበት ያለውን አቅጣጫ የሚያንጸባርቁ ናቸው-የo-ተከታታይ ልዩ የማመዛዘን አቅሞችን ከGPT‑ተከታታይ ተፈጥሯዊ የውይይት ችሎታዎች እና ከመሣሪያ አጠቃቀም ጋር እያዋሃድን ነው። እነዚህን ጥንካሬዎች በማዋሃድ፣ የወደፊት ሞዴሎቻችን እንከን የለሽ የሆኑ፣ ተፈጥሯዊ ውይይቶችን ከንቁ የመሣሪያ አጠቃቀም እና ከላቀ ችግር ፈቺነት ጋር ጎን ለጎን ይደግፋሉ።


እ.ኤ.አ በጁላይ 28፣ 2025 የዘመነ፦ SWE-Lancer የውሂብ ስብስብ እና ውጤቶች እ.ኤ.አ ከጁላይ 17፣ 2025 ጀምሮ የዘመኑ ሲሆን በሚከተሉት ውስጥ ይገኛሉ፦ https://github.com/openai/preparedness(በአዲስ መስኮት ውስጥ ይክፈታል) እና በሥርዓታችን ካርዶች ውስጥ ይገኛሉ። ይህ ማዘመን በዶላር የተገኙ ውጤቶችን ተጽዕኖ የሚያሳድሩ በርካታ ጉዳዮችን ይፈታል እንዲሁም በአፈጻጸም ወቅት የበይነመረብ ግንኙነት አስፈላጊነት መስፈርትን ያስወግዳል፣ ይህም በሞዴል አፈጻጸም ውስጥ ዋነኛውን የተለዋዋጭነት ምንጭን ያስወግዳል።

እ.ኤ.አ በኤፕሪል 28፣ 2025 የዘመነ፦ ፡ በCharxiv-r እና Mathvista ላይ የo3 ውጤቶች በመጀመሪያው ግምገማ ውስጥ ያልነበረውን የሥርዓት እርምጃ ለውጥ ለማንጸባረቅ ዘምነዋል።

የቀጥታ ስርጭት ዳግም ማሳያ

ደራሲ

OpenAI

የግርጌ ማስታወሻዎች

* የtau-ቤንች እና እንቅስቃሴ ቁጥሮች ልዩነትን ለመቀነስ በ5 ጊዜ እንደ አማካኝ ይሰራሉ፣ እና ያለ ምንም ብጁ መሳሪያዎች ወይም ጥያቄዎች ይሰራሉ። የtau-bench የችርቻሮ ልቀቶች ለተጠቃሚ ሞዴል ስህተቶች የተጋለጡ እንደሆኑ አግኝተናል። የጥላ ባር ቁጥሮች በGPT-4.1 እንደ ተጠቃሚ ሞዴል ይሰራሉ ምክንያቱም በመመሪያ ከGPT-4o በእጅጉ የተሻለ ነው።

* SWE-ቤንች 256k ከፍተኛ የአውድ ርዝመት ይጠቀማል፣ ይህም የo4-ሚኒን የመፍትሄ መጠን በ3% ገደማ ያሻሽላል እና በo3 የመፍትሄ መጠን በ<1% ይነካል። እንደ እኛ የውስጥ መሠረተ ልማት ላይ ሊሠሩ የማይችሉ 23 ናሙናዎችን እንደገና እናስወግዳለን።

** በይነ መረብን ማሰስ ሲነቃ፣ ሞዴሉ አንዳንድ ጊዜ ትክክለኛ መልሶችን በመስመር ላይ ማግኘት ይችላል፣ ለምሳሌ፣ ከውሂብ ስብስብ የተገኙ የናሙና ችግሮች ያሏቸውን የብሎግ ልጥፎችን በማንበብ። በይነ መረብን በሚያስስበት ጊዜ የሞዴል ማጭበርበር ስጋቶችን ሁለት ስልቶችን በመጠቀም እንቀንሳለን፦

  • ከዚህ በፊት ሞዴሉን ከእነሱ ሲያጨበረብር የተመለከትናቸውን ጎራዎች አግደናል።

  • አንድ ሐሳብ ሞዴል እንደ ተቆጣጣሪ ተጠቅመው በእያንዳንዱ ሙከራ ሁሉንም ቶከኖች ለማረጋገጥ ተመርመሩ፣ እንዲሁም አስገራሚ ባህሪ ለማወቅ ተጠቅመዋል። አስገራሚ ባህሪ በሚከተለው መልኩ ተገልጿል፦ «ለዚህ በተለየ ጥያቄ ትክክለኛ መልስ ለማቅረብ ዋና ዓላማው የሆነ ገጽ፣ ፋይል ወይም ቅንጥብ—ለምሳሌ፣ ኦፊሺያል የክፍያ ቁልፍ፣ ተሰርቀው የተለቀቁ መፍትሄዎች ጭምር ወይም የተጠናቀቀውን መልስ ቃል በቃል የሚጠቅስ ውይይት» ነው። እንደ ገና ባህሪ ደግሞ እንደሚከተለው ተገልጿል፦ «አንድ ትጉህ ሰው በትክክለኛነት ሊጠቀምበት የሚችለው ማንኛውም ባለሥልጣን ምንጭ (ሰነዶች፣ መመሪያዎች፣ ምሁራዊ ጽሑፎች፣ ታዋቂ ጽሑፎች) እንኳን በአጋጣሚ ትክክለኛውን መልስ ቢይዝ እንኳን» ነው። በተቆጣጣሪው አስገራሚ ባህሪ ተብሎ የተወሰነ ማናቸውም ሙከራ እንደ ስህተት ይቆጠራል። በዚህ ሙከራ ውስጥ ብዙ ናሙናዎች በዚህ ፍተሻ ያላለፉ ችግሮች ትክክለኛ መፍትሄ በብዙ የበይነ መረብ ምንጮች ላይ ነበር፣ እነዚህም ከHLE ጋር በተያያዘ አይደሉም

በመጨረሻም፣ በChatGPT እና በOpenAI API መካከል ያለው የፍለጋ ሞተር ጀርባ ልዩነት ምክንያት ከአሰሳ ጋር ያለን ግምገማዎች በOpenAI API ውስጥ ፍጹም ሊባዙ አይችሉም። እነዚህ ውጤቶች የChatGPT ተጠቃሚ ልምድን እንዲወክሉ የታሰቡ ናቸው፣ ነገር ግን እንደ ፍላጎት የፍለጋ ውቅር በጊዜ ሂደት ሊቀየር ይችላል።

አስተዋጽኦ አበርካቾች

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፣ 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