Maxaa dhacay
Sida laga soo xigtay Northeast Times, Prime Intellect's NanoGPT Speedrun ayaa qiimeeyay 18 nooc oo AI ah oo ku saabsan 153 imtixaan oo u baahday nidaam kasta si uu u wanaajiyo macalinka nooca luuqada yar iyada oo aan la helin caawimaad bini'aadan. Claude Opus 5 wuxuu dhamaystiray 81.7% ee bartilmaameedka, waxaa ku xiga Kimi K3 oo helay 52.2% iyo GPT-5.6 Sol oo helay 35.9%. Warbixintu waxay sheegtay in Prime Intellect uu sidoo kale daabacay 41 raadraac raadraac oo muujinaya sida moodooyinka loo isticmaalo aaladaha, xusuusta iyo API-yada dibadda. Natiijooyinka bartilmaameedka si madax banaan looma xaqiijin halkan.
Northeast Times reports that Prime Intellect’s NanoGPT Speedrun placed 18 AI models through 153 independent tests. Each test asked a model to optimize a small language-model trainer autonomously, without human assistance. The source presents this as a measure of sustained coding and debugging ability rather than a simple code-generation exercise. The supplied report links to Prime Intellect’s page, but the benchmark methodology and raw results are not independently confirmed in this evaluation.
The reported ranking was sharply uneven. Northeast Times says Claude Opus 5 completed 81.7% of the suite, while Kimi K3 completed 52.2% and GPT-5.6 Sol completed 35.9%. Claude Sonnet 5, GPT-5.6 Luna and Grok 4.5 reportedly fell in the 20% to 26% range. DeepSeek V4 Pro, Muse Spark 1.2 and GPT-5.5 were reported below 15%. These figures are attributed to the Northeast Times report and should not be treated as independently audited scores.
The source says the evaluation tracked more than final completion rates. Northeast Times reports that stronger systems reached accuracy thresholds with fewer optimization steps and smaller memory footprints, while weaker systems often ran longer without meaningful improvement. The report attributes this interpretation to Hyper.ai, which described a capability gap involving multi-step code generation and error recovery. The supplied material does not provide the underlying measurements, confidence intervals or detailed task-by-task results.
Northeast Times also reports that Prime Intellect published 41 fully traced agent trajectories. These records allegedly show tool calls, memory allocation patterns, error-handling routines, scratchpad reasoning and interactions with external APIs. The source says top-performing systems used structured subagent delegation and systematic tool invocation, while weaker systems sometimes entered recursive loops or stopped improving prematurely. The article does not establish whether all 18 models received identical scaffolding, tool access or budgets.
Faahfaahinta isha: northeasttimes.com ↗
Maxay muhiim u tahay
Faafida la soo sheegay waxay soo jeedinaysaa in xulashada moodalku ay si wax ku ool ah u saameyn karto isku halaynta codaynta qulqulka shaqada. Natiijooyinka waxay sidoo kale tilmaamayaan muhiimada naqshadeynta wakiilka, isticmaalka qalabka iyo soo kabashada qaladka, ma aha oo kaliya cabbirka qaabka ama awoodda cinwaanka. Ururada tixgelinaya dib-u-soo-noolaynta tooska ah, abuuritaanka kaabayaasha ama is-dhexgalka joogtada ah, qiimaynta kood la soo saari karo ayaa laga yaabaa inay ka macluumaad badan tahay bandhigyada go'doonsan. Natiijooyinka ayaa weli xaddidaya naqshadaynta hawsha bartilmaameedka iyo la'aanta xaqiijinta madaxbannaan ee walxaha la keenay.
The reported results matter because autonomous coding systems are increasingly evaluated by whether they can complete multi-step work, not merely produce plausible snippets. A model that can recover from errors, manage tools and continue toward a target may be more useful than one that performs well on isolated prompts. Northeast Times connects the to enterprise uses such as automated refactoring, continuous integration and infrastructure generation, although the article does not document specific deployments or measured business outcomes.
Baaxadda faraqa waxqabadka ee la soo sheegay ayaa caqabad ku ah malo-awaalka ah in kala duwanaanshaha kood-samaynta ay yihiin kuwo yar. Marka la eego tirooyinka ay soo xigtay Northeast Times, Claude Opus 5 waxay dhamaystirtay hawlo aad uga badan marka loo eego nidaamyada darajada xiga iyo in ka badan shan jeer heerka heerka ugu waxqabadka hooseeya. Isbarbardhiggaas ayaa suurtagal u ah in uu muhiim u yahay ururrada dooranaya qaabka, laakiin waxa ay ahaanaysaa mid gaar ah bartilmaameedkan. Ma cadda in hal nooc uu si caalami ah uga wanaagsan yahay luqadaha barnaamijyada, kaydinta, hawlaha amniga ama deegaanka wax soo saarka.
Bartilmaameedka ayaa sidoo kale iftiiminaya farqiga u dhexeeya awoodda moodeelka iyo qaabeynta nidaamka. Waqooyi-bari Times waxay leedahay hawl-wadeennada ugu xoogga badan waxay ku tiirsanaayeen ergada habaysan iyo adeegsiga aaladaha, halka nidaamyada daciifka ahi ay dhici karaan inay goor hore soo xidhmaan ama isu yimaadaan. Haddii ay sax tahay, taasi waxay la macno tahay in buundada moodeelka qayb ahaan ay ka tarjumayso suunka wakiilka, dardargelinta, agabka iyo xadka ilaha. Maqaalku ma daaha ka qaadayo faahfaahin ku filan qaabeynta si loo go'aamiyo inta ka mid ah darajooyinka ka yimid moodooyinka hoose iyo inta ka timid habayntooda hawlgelinta.
Daah-furnaantu waxay ka dhigi kartaa qiimaynta mid ka faa'iido badan hal guddi haddii horumariyeyaasha dibadda ay baadhi karaan oo ay soo saari karaan raadadka. Northeast Times waxay ku qeexday 41-da raadraac inay yihiin cadaymo faahfaahsan oo aan caadi ahayn oo ku saabsan sida moodalku u shaqeeyaan hawlaha codaynta. Diiwaanada noocaan ah ayaa kaa caawin kara in la aqoonsado hababka guul dareysiga iyo wanaajinta baaritaanka. Si kastaba ha ahaatee, raadadka daabacaadda lafteedu ma caddaynayso in bartilmaameedku yahay wakiil, in hawshu aan lagu hagaajin nidaamyo gaar ah, ama in natiijadu ay ka soo baxayso qolka la soo sheegay.
Farsamaynta Is-dhexgalka: Sida Dhabta Ay U Shaqeyso
U baadh tignoolajiyada hoose ee ka dambeeya horumarkan si isdhexgal leh.
Which component of an AI application is the machine-learning model itself?
Maxaa la daawan doona xiga
The key questions are whether the NanoGPT Speedrun tasks represent real software-engineering work, whether the ranking holds across other codebases and languages, and whether the results can be reproduced by outside evaluators. Developers should examine the published traces and test models on their own repositories before treating the scores as deployment guidance. Future evaluations should report costs, latency, failure severity and human-review requirements alongside completion rates.
Arrinta ugu horreysa ee la daawado waa taransiinta. Isha ayaa sheegaysa in Prime Intellect uu sameeyay 41 wakiil oo raadraac ah, laakiin ma sheegayso in hawsha dhameystiran, koodka dhibcaha, noocyada moodeelka, dardargelinta, ogolaanshaha qalabka iyo xaddidaadaha kheyraadka ay yihiin dad weyne. Dib u socodsiinta madaxa banaan ee la isticmaalayo isla shuruudo ayaa gacan ka geysan doonta in la ogaado in darajada la soo sheegay ay tahay mid degan halkii ay ka ahaan lahayd farshaxan hal habayn oo qiimayn ah.
Arrinka labaad waa ansaxnimada dibadda. Hagaajinta macalinka-qaabka yar ee luuqada waxa laga yaabaa inay tijaabiso xirfadaha waxtarka leh ee ciribtirka, tijaabinta iyo soo noqnoqoshada joogtada ah, laakiin la mid ma aha ilaalinta kaydka wax soo saarka weyn. Isbarbardhigga mustaqbalka waa in lagu daro imtixaanada dib u eegista koodka, maaraynta ku tiirsanaanta, dayacanka amniga, dukumentiyada, socdaalka xogta iyo isbeddelada kaydka ee socda. Warbixinta la keenay ma tusinayso sida hawlaha la qiimeeyay u maabeeyeen goobahaas.
Kharashka iyo waxqabadka hawlqabadku sidoo kale waxay u baahan yihiin baadhitaan. Northeast Times waxay ka warbixisaa kala duwanaanshaha tillaabooyinka hagaajinta iyo raadadka xusuusta, laakiin ma bixiso qiimayaal, daahitaan, wadarta isticmaalka calaamada, isticmaalka tamarta ama kharashyada soo kabashada. Qaabka leh heerka dhamaystirka sare waxa laga yaabaa in aanu noqon doorashada ganacsi ee ugu wanagsan haddii uu aad qaali u yahay ama uu u baahan yahay dib u eegis balaadhan oo biniaadmi ah. Tallaabooyinkaas wax ku oolka ah waa inay la socdaan buundooyinka sabuuradaha mustaqbalka.
Ugu dambayntii, ururadu waa in ay natiijada u daaweeyaan calaamad baadhiseed halkii ay ka ahaan lahaayeen dammaanad geyn. Northeast Times waxay soo xiganeysaa naqshadeeye ganacsi oo yiri isdhexgalku waa arinta udub dhexaadka, laakiin aragtidaas waa faallooyin halkii ay ka ahaan lahayd caddayn madaxbannaan. Kooxuhu waa in ay tijaabiyaan nidaamyada musharraxiinta iyaga oo ka soo horjeeda kaydkooda, qeexaan hababka guul darada ee la aqbali karo oo ay u baahan yihiin dib u eegis ka hor inta uusan koodka gaarin wax soo saarka. Gabagabada bartilmaameedka ayaa laga yaabaa inay isbedesho marka moodooyinka, qallafsanaanta iyo hawlaha qiimayntu ay horumaraan.