benchmark

approfondimento
Come misurare la qualità di un output AI: eval e calibrazione del modello giudice

09 Set 2026
<img width="768" height="432" src="https://img.ai4business.it/wp-content/uploads/2026/09/artificial-analysis-index-768x432.png" class="card-responsive-img fit-img" alt="artificial analysis index" sizes="(min-width: 1348px) 394px, (min-width: 992px) calc(26.19vw + 22px), (min-width: 600px) calc(50vw - 58px), calc(100vw - 34px)" decoding="async" srcset="https://img.ai4business.it/wp-content/uploads/2026/09/artificial-analysis-index-768x432.png 768w, https://img.ai4business.it/wp-content/uploads/2026/09/artificial-analysis-index-lq.png 30w, https://img.ai4business.it/wp-content/uploads/2026/09/artificial-analysis-index-156x88-1.png 156w, https://img.ai4business.it/wp-content/uploads/2026/09/artificial-analysis-index-1536x864.png 1536w, https://img.ai4business.it/wp-content/uploads/2026/09/artificial-analysis-index-1280x720.png 1280w, https://img.ai4business.it/wp-content/uploads/2026/09/artificial-analysis-index-640x360.png 640w, https://img.ai4business.it/wp-content/uploads/2026/09/artificial-analysis-index-1040x585.png 1040w, https://img.ai4business.it/wp-content/uploads/2026/09/artificial-analysis-index-1024x576.png 1024w, https://img.ai4business.it/wp-content/uploads/2026/09/artificial-analysis-index.png 1920w" data-eio="l" />

approfondimento
Benchmark AI, l'Intelligence Index riscritto due volte in quattro giorni
09 Set 2026
di Fabio Lalli
Condividi il post
Condividi



<img width="768" height="433" src="https://img.ai4business.it/wp-content/uploads/2026/09/image-31-e1788517264520-768x433.png" class="card-responsive-img fit-img" alt="Gemini 3.8 Flash: Google guadagna terreno nel coding" sizes="(min-width: 1348px) 394px, (min-width: 992px) calc(26.19vw + 22px), (min-width: 600px) calc(50vw - 58px), calc(100vw - 34px)" decoding="async" srcset="https://img.ai4business.it/wp-content/uploads/2026/09/image-31-e1788517264520-768x433.png 768w, https://img.ai4business.it/wp-content/uploads/2026/09/image-31-e1788517264520-1278x720.png 1278w, https://img.ai4business.it/wp-content/uploads/2026/09/image-31-e1788517264520-156x88.png 156w, https://img.ai4business.it/wp-content/uploads/2026/09/image-31-e1788517264520-1918x1080.png 1918w, https://img.ai4business.it/wp-content/uploads/2026/09/image-31-e1788517264520-1536x865.png 1536w, https://img.ai4business.it/wp-content/uploads/2026/09/image-31-e1788517264520-2048x1153.png 2048w, https://img.ai4business.it/wp-content/uploads/2026/09/image-31-e1788517264520-640x360.png 640w, https://img.ai4business.it/wp-content/uploads/2026/09/image-31-e1788517264520-1024x577.png 1024w, https://img.ai4business.it/wp-content/uploads/2026/09/image-31-e1788517264520-1040x585.png 1040w" data-eio="l" />

approfondimento
Gemini 3.8 Flash: Google guadagna terreno nel coding
04 Set 2026
di Fabio Lalli
Condividi il post
Condividi



<img width="768" height="434" src="https://img.ai4business.it/wp-content/uploads/2026/08/image-60-e1787130570430-768x434.png" class="card-responsive-img fit-img" alt="GLM-5.3 cybersecurity" sizes="(min-width: 1348px) 394px, (min-width: 992px) calc(26.19vw + 22px), (min-width: 600px) calc(50vw - 58px), calc(100vw - 34px)" decoding="async" srcset="https://img.ai4business.it/wp-content/uploads/2026/08/image-60-e1787130570430-768x434.png 768w, https://img.ai4business.it/wp-content/uploads/2026/08/image-60-e1787130570430-156x88.png 156w, https://img.ai4business.it/wp-content/uploads/2026/08/image-60-e1787130570430-1024x578.png 1024w, https://img.ai4business.it/wp-content/uploads/2026/08/image-60-e1787130570430.png 1240w" data-eio="l" />

ai agentica
Z.ai lancia GLM-5.3: cresce il rischio cyber dell’AI
19 Ago 2026
di Pierluigi Sandonnini
Condividi il post
Condividi



<img width="768" height="433" src="https://img.ai4business.it/wp-content/uploads/2026/08/image-47-768x433.png" class="card-responsive-img fit-img" alt="Gemini 3.7 Flash" sizes="(min-width: 1348px) 394px, (min-width: 992px) calc(26.19vw + 22px), (min-width: 600px) calc(50vw - 58px), calc(100vw - 34px)" decoding="async" srcset="https://img.ai4business.it/wp-content/uploads/2026/08/image-47-768x433.png 768w, https://img.ai4business.it/wp-content/uploads/2026/08/image-47-1277x720.png 1277w, https://img.ai4business.it/wp-content/uploads/2026/08/image-47-156x88.png 156w, https://img.ai4business.it/wp-content/uploads/2026/08/image-47-1536x866.png 1536w, https://img.ai4business.it/wp-content/uploads/2026/08/image-47-640x360.png 640w, https://img.ai4business.it/wp-content/uploads/2026/08/image-47-1024x577.png 1024w, https://img.ai4business.it/wp-content/uploads/2026/08/image-47-1040x585.png 1040w, https://img.ai4business.it/wp-content/uploads/2026/08/image-47-lq.png 30w, https://img.ai4business.it/wp-content/uploads/2026/08/image-47.png 1600w" data-eio="l" />

AI agentica
Google Gemini 3.7 Flash: più potenza, meno costi
17 Ago 2026
di Pierluigi Sandonnini
Condividi il post
Condividi



<img width="433" height="243" src="https://img.ai4business.it/wp-content/uploads/2026/06/image-91-e1785488997323.jpg" class="card-responsive-img fit-img" alt="GPT-5.6" sizes="(min-width: 1348px) 394px, (min-width: 992px) calc(26.19vw + 22px), (min-width: 600px) calc(50vw - 58px), calc(100vw - 34px)" decoding="async" srcset="https://img.ai4business.it/wp-content/uploads/2026/06/image-91-e1785488997323.jpg 433w, https://img.ai4business.it/wp-content/uploads/2026/06/image-91-e1785488997323-156x88.jpg 156w" data-eio="l" />

analisi
GPT-5.6, Luna costa l'80% in meno, ma il costo per task non scende
31 Lug 2026
di Fabio Lalli
Condividi il post
Condividi



<img width="768" height="359" src="https://img.ai4business.it/wp-content/uploads/2026/07/image-93-768x359.png" class="card-responsive-img fit-img" alt="Humanity&#039;s Last Exam" sizes="(min-width: 1348px) 394px, (min-width: 992px) calc(26.19vw + 22px), (min-width: 600px) calc(50vw - 58px), calc(100vw - 34px)" decoding="async" srcset="https://img.ai4business.it/wp-content/uploads/2026/07/image-93-768x359.png 768w, https://img.ai4business.it/wp-content/uploads/2026/07/image-93-1280x599.png 1280w, https://img.ai4business.it/wp-content/uploads/2026/07/image-93-156x73.png 156w, https://img.ai4business.it/wp-content/uploads/2026/07/image-93-1024x479.png 1024w, https://img.ai4business.it/wp-content/uploads/2026/07/image-93-lq.png 30w, https://img.ai4business.it/wp-content/uploads/2026/07/image-93.png 1435w" data-eio="l" />

approfondimento
Humanity's Last Exam, cos'è e perché è uno dei benchmark più quotati
22 Lug 2026
di Pierluigi Sandonnini
Condividi il post
Condividi



<img width="768" height="468" src="https://img.ai4business.it/wp-content/uploads/2026/06/Models-Struggle-When-Science-Moves-Beyond-Text-768x468.png" class="card-responsive-img fit-img" alt="LifeSciBench" sizes="(min-width: 1348px) 394px, (min-width: 992px) calc(26.19vw + 22px), (min-width: 600px) calc(50vw - 58px), calc(100vw - 34px)" decoding="async" srcset="https://img.ai4business.it/wp-content/uploads/2026/06/Models-Struggle-When-Science-Moves-Beyond-Text-768x468.png 768w, https://img.ai4business.it/wp-content/uploads/2026/06/Models-Struggle-When-Science-Moves-Beyond-Text-1183x720.png 1183w, https://img.ai4business.it/wp-content/uploads/2026/06/Models-Struggle-When-Science-Moves-Beyond-Text-1774x1080.png 1774w, https://img.ai4business.it/wp-content/uploads/2026/06/Models-Struggle-When-Science-Moves-Beyond-Text-1536x935.png 1536w, https://img.ai4business.it/wp-content/uploads/2026/06/Models-Struggle-When-Science-Moves-Beyond-Text-145x88.png 145w, https://img.ai4business.it/wp-content/uploads/2026/06/Models-Struggle-When-Science-Moves-Beyond-Text-1024x623.png 1024w, https://img.ai4business.it/wp-content/uploads/2026/06/Models-Struggle-When-Science-Moves-Beyond-Text-lq.png 30w, https://img.ai4business.it/wp-content/uploads/2026/06/Models-Struggle-When-Science-Moves-Beyond-Text.png 2040w" data-eio="l" />

scenari
LifeSciBench, il benchmark che misura l'AI sulla ricerca biologica reale
18 Giu 2026
di Fabio Lalli
Condividi il post
Condividi



<img width="688" height="458" src="https://img.ai4business.it/wp-content/uploads/2026/04/image-117.png" class="card-responsive-img fit-img" alt="gpt 5.5" sizes="(min-width: 1348px) 394px, (min-width: 992px) calc(26.19vw + 22px), (min-width: 600px) calc(50vw - 58px), calc(100vw - 34px)" decoding="async" srcset="https://img.ai4business.it/wp-content/uploads/2026/04/image-117.png 688w, https://img.ai4business.it/wp-content/uploads/2026/04/image-117-132x88.png 132w, https://img.ai4business.it/wp-content/uploads/2026/04/image-117-lq.png 30w" data-eio="l" />

analisi
Modelli AI avanzati: valutazioni di terze parti più trasparenti per misurare rischi e capacità
01 Giu 2026
di Pierluigi Sandonnini
Condividi il post
Condividi