#geminipro — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #geminipro, aggregated by home.social.
-
Gemini «olvida» mucho antes de lo que Google promete: el millón de tokens no aplica al chat
Usuarios de los planes pagos AI Pro y Ultra denuncian que el chatbot comienza a perder el hilo de la conversación después de apenas 25-30 mensajes, muy lejos del límite de un millón de tokens que Google publicita. La diferencia entre la ventana de contexto del modelo y la del chat nunca se comunica con claridad (Fuente AndroidAutorithy).
Google vende sus planes pagos de Gemini con una promesa concreta: una ventana de contexto de hasta un millón de tokens, equivalente a 1.500 páginas de texto o 30.000 líneas de código. El problema es que esa cifra no describe lo que le pasa al usuario en una conversación real.
Usuarios en X y Reddit denunciaron que, si bien los servidores de Gemini pueden efectivamente ingerir un archivo estático masivo en el primer prompt, la memoria conversacional activa —el contexto dinámico del chat— parece estar severamente limitada, cayendo a un tope aproximado de 16.000 tokens, equivalente a unos 25 o 30 mensajes promedio. El resultado es que el modelo sufre de «amnesia» dentro de la misma sesión de chat, olvidando por completo instrucciones anteriores, bloques de código o restricciones que el usuario había establecido al inicio de la conversación.
La distinción técnica que Google no comunica con suficiente claridad es la diferencia entre la ventana de contexto del modelo y la del chat. En palabras del usuario de X @Soso_fun_yt, mientras el backend puede procesar archivos de gran tamaño en forma estática, la memoria dinámica de la conversación está embotellada en un límite mucho menor, lo que provoca el olvido progresivo. Algunos usuarios señalaron que la plataforma AI Studio sí ofrece la ventana de contexto correcta, pero esa no es la herramienta que usa la mayoría de los suscriptores.
La analogía que propone el artículo lo dice todo: es como si tu proveedor de internet anunciara una línea de 1 Gbps en su sitio web, sin mencionar en ningún lugar destacado que la velocidad de subida es de apenas 50 Mbps. Google sí publica información técnica sobre tokens de entrada y salida en su documentación para desarrolladores, pero esa información no llega al usuario promedio que paga su suscripción esperando lo que se le prometió.
Android Authority consultó a Google sobre la discrepancia entre la ventana de contexto del modelo y la del chat, y sobre si planea ofrecer información más prominente al respecto. La compañía no respondió al momento de publicación. Mientras tanto, quienes usan Gemini para proyectos largos o conversaciones técnicas extendidas deberían saber que el millón de tokens es, por ahora, más un horizonte teórico que una realidad práctica de uso diario.
#AIStudio #AIUltra #chatbot #contextwindow #gemini #GeminiAI #GeminiPro #google #GoogleAI #IA #InteligenciaArtificial #PORTADA #Suscripciones #tokens #Transparencia -
Los nuevos límites de Gemini fallaron y Google tuvo que dar marcha atrás
Apenas una semana después de presentar en el Google I/O 2026 su nuevo sistema de límites de uso basado en poder de cómputo, Google se vio obligado a aplicar parches de urgencia tras una avalancha de quejas: usuarios suscriptores pagos reportaban que un solo prompt de generación de video consumía toda su cuota antes de que el video siquiera terminara de generarse.
https://twitter.com/joshwoodward/status/2060171610922058142
Google tuvo una semana complicada con Gemini. En el Google I/O 2026, la compañía presentó un nuevo sistema de límites de uso basado en poder de cómputo, reemplazando el modelo anterior que contabilizaba cada prompt por igual. La idea sonaba razonable sobre el papel: medir cuánto procesamiento consume cada solicitud en lugar de contar interacciones individuales. Luego los usuarios lo pusieron a prueba, y el rechazo fue inmediato.
Los casos más extremos resultaron difíciles de ignorar. Un suscriptor del plan Google AI Pro publicó un video como prueba mostrando cómo un único intento de generación de avatar consumió toda su cuota de cinco horas antes de que el video terminara de generarse. No fue el único: las primeras reacciones al cambio fueron mayoritariamente negativas, con usuarios cuestionando incluso la forma en que Google comunicó el nuevo sistema, que comparaba el plan Pro contra el nivel gratuito en lugar de explicar qué ofrecía en relación a lo que el propio plan Pro ofrecía antes.
El nuevo esquema, que entró en vigor el 20 de mayo, funciona de la siguiente manera: los límites se renuevan cada cinco horas hasta alcanzar un tope semanal, y el cálculo de consumo toma en cuenta la complejidad del prompt, las funciones utilizadas y la extensión de la conversación. Las tareas que más cuota consumen incluyen generación de imágenes, videos y música, Deep Research, el modelo Pro y las funciones de pensamiento extendido como Deep Think.
La respuesta de Google llegó rápido. Josh Woodward, líder del equipo de Gemini, anunció públicamente que Google está aplicando un tope al máximo de cuota que un solo prompt puede consumir al usar Gemini 3.1 Pro, de modo que una solicitud pesada no pueda agotar el presupuesto completo de una sesión. Además, Google confirmó que en el futuro los usuarios de la app de Gemini podrán comprar créditos de IA de pago por uso para continuar trabajando una vez alcanzado el límite, sin tener que esperar el reinicio de la cuota.
El contexto de fondo es relevante: el cambio llega menos de un mes después de que GitHub reformulara su plan Copilot, también abandonando un modelo de solicitudes fijas para adoptar «AI Credits» basados en tokens realmente utilizados. La tendencia es clara: los planes de tarifa plana para IA generativa están llegando a su límite natural frente a la creciente demanda de funciones computacionalmente intensivas. La pregunta que queda es si los usuarios pagarán más o simplemente utilizarán menos.
#AIUltra #cómputo #DeepResearch #gemini #GeminiPro #google #GoogleAI #GoogleIO2026 #IA #InteligenciaArtificial #JoshWoodward #limitesdeuso #PORTADA #Suscripciones #tecnologia -
#Google's new #GeminiPro model has record benchmark scores again: Gemini 3.1 Pro promises a Google LLM capable of handling more complex forms of work. The model has achieved record scores on key benchmarks, further intensifying the #AIrace among #techgiants. https://techcrunch.com/2026/02/19/googles-new-gemini-pro-model-has-record-benchmark-scores-again/?eicker.news #tech #media #news
-
#Google's new #GeminiPro model has record benchmark scores again: Gemini 3.1 Pro promises a Google LLM capable of handling more complex forms of work. The model has achieved record scores on key benchmarks, further intensifying the #AIrace among #techgiants. https://techcrunch.com/2026/02/19/googles-new-gemini-pro-model-has-record-benchmark-scores-again/?eicker.news #tech #media #news
-
Gemini 3 Pro tăng mức tin cậy lên 69% trong thử nghiệm mù (tăng từ 16% so với phiên bản trước), dẫn đầu 3/4 hạng mục: hiệu suất, tương tác và an toàn. Dù bị DeepSeek V3 vượt về phong cách giao tiếp, nó vẫn giành 5 lần lựa chọn ưu tiên của người dùng. Đánh giá HUMAINE nhấn mạnh tính nhất quán qua 22 nhóm nhân khẩu học, cảnh báo doanh nghiệp nên ưu tiên mô hình phù hợp nhu cầu và đối tượng cụ thể.
#AI #GeminiPro #ĐánhGiáAI #CôngNghệ #TrustworthyAI
#ThửNghiệmMù #ĐaDạngNgườiDùng #TriTueNhanTao -
Interesting how Poetiq (company) can improve on the performance of the standard Gemini 3.0 Pro model by adding refinements and tricks. It leads to a 9% improvement on the ARC-AGI-2 Benchmark.
https://poetiq.ai/posts/arcagi_verified/
#AI #Poetiq #ArcAGI2 #Google #GeminiPro -
Interesting how Poetiq (company) can improve on the performance of the standard Gemini 3.0 Pro model by adding refinements and tricks. It leads to a 9% improvement on the ARC-AGI-2 Benchmark.
https://poetiq.ai/posts/arcagi_verified/
#AI #Poetiq #ArcAGI2 #Google #GeminiPro -
Google Gemini May Get Two New Subscription Tiers #ai #artificalintelligence #artificialintelligence #gemini #geminiadvanced #geminipro #geminiultra #google #oneaipremium
https://www.lowyat.net/2025/350341/google-gemini-may-get-two-new-subscription-tiers/
-
Google Gemini May Get Two New Subscription Tiers #ai #artificalintelligence #artificialintelligence #gemini #geminiadvanced #geminipro #geminiultra #google #oneaipremium
https://www.lowyat.net/2025/350341/google-gemini-may-get-two-new-subscription-tiers/
-
Gemini 2.5 Pro Appears to Be the First AI Model to Fully Understand PDF Layouts, Enabling Precise Citations
#AI #GoogleAI #Gemini #GeminiPro #LLM #PDF #MultimodalAI #AICitations #Alphabet #Google #GenAI
-
Gemini 2.5 Pro Appears to Be the First AI Model to Fully Understand PDF Layouts, Enabling Precise Citations
#AI #GoogleAI #Gemini #GeminiPro #LLM #PDF #MultimodalAI #AICitations #Alphabet #Google #GenAI
-
🤖 AI
🔴 Google’s Gemini 2.5 Pro Raises the Bar
🔸 Beats o3-mini, Grok & DeepSeek in chatbot rankings.
🔸 Fast, accurate & intuitive on coding, math, and science.
🔸 Context window 5x larger than rivals—2M token upgrade coming.
🔸 Now in mobile & web; 50 free messages/day before pricing update.#Gemini25 #GoogleAI #Chatbots #AIModels #LLM #OpenAI #GeminiPro #VertexAI
-
🤖 AI
🔴 Google’s Gemini 2.5 Pro Raises the Bar
🔸 Beats o3-mini, Grok & DeepSeek in chatbot rankings.
🔸 Fast, accurate & intuitive on coding, math, and science.
🔸 Context window 5x larger than rivals—2M token upgrade coming.
🔸 Now in mobile & web; 50 free messages/day before pricing update.#Gemini25 #GoogleAI #Chatbots #AIModels #LLM #OpenAI #GeminiPro #VertexAI
-
#Phi4 from #Microsoft Sets New Benchmark in Small Language Models 🚀
🧮 #Phi4 achieves 91.8% accuracy on November 2024 #AMC tests, surpassing #GeminiPro and other larger #LLM models
🧵 ↓ #ai
-
I am using both GPT-4o and NotebookLM, i have the impression that NotebookLM is much better at summarizing and analyzing texts, i am wondering whether that's because of NotebookLM implementation aspects or because of that Gemini Pro 1.5 is better at this task then GPT-4o ? 🤔
#AI #NotebookLM #Geminipro #ChatGPT #openai #google #text #summarizing #analysing -
I am using both GPT-4o and NotebookLM, i have the impression that NotebookLM is much better at summarizing and analyzing texts, i am wondering whether that's because of NotebookLM implementation aspects or because of that Gemini Pro 1.5 is better at this task then GPT-4o ? 🤔
#AI #NotebookLM #Geminipro #ChatGPT #openai #google #text #summarizing #analysing -
ChatGPT-o1 was able to describe this 6502 assembly language function:
```
Here is a 6502 assembly language routine. The ROM_COUT routine prints the character stored in the A register. Generate a comment for this routine describing what it is for, and any inputs and outputs. Also indicate which registers are used and would have to be saved before calling the routine.routine:
LDY #$00
LDA ($B0),Y
TAX.a:
DEX
BPL .b
RTS.b:
INY
LDA ($B0),Y
JSR ROM_COUT
JMP .a
```The response seems accurate to me:
```
; Routine: Print Length-Prefixed String from Memory Address in $B0/$B1
;
; Description:
; This routine outputs a length-prefixed string to the console using the ROM_COUT subroutine.
; The string is stored in memory at the address specified by the zero-page pointers $B0 (low byte) and $B1 (high byte).
; The first byte at this address represents the length 'n' of the string, followed by 'n' characters to be printed.
;
; Inputs:
; - $B0/$B1: Zero-page pointers containing the 16-bit address of the length-prefixed string.
;
; Outputs:
; - The string is printed to the console, character by character, via the ROM_COUT routine.
;
; Registers Used (and modified):
; - A (Accumulator): Used to load string length and characters.
; - X: Used as a counter to keep track of remaining characters to print.
; - Y: Used as an index to traverse the string in memory.
;
; Note:
; - If the calling code requires the values in registers A, X, or Y after this routine executes,
; it should save and restore them before and after the call, respectively.
```As a second data point, Gemini got it wrong, claiming that the string had to be null-terminated, and that the string had to be located at $B0.
Claude (Sonnet) also claimed the string had to be null-terminated, but did see that the string's address was at $B0.
-
ChatGPT-o1 was able to describe this 6502 assembly language function:
```
Here is a 6502 assembly language routine. The ROM_COUT routine prints the character stored in the A register. Generate a comment for this routine describing what it is for, and any inputs and outputs. Also indicate which registers are used and would have to be saved before calling the routine.routine:
LDY #$00
LDA ($B0),Y
TAX.a:
DEX
BPL .b
RTS.b:
INY
LDA ($B0),Y
JSR ROM_COUT
JMP .a
```The response seems accurate to me:
```
; Routine: Print Length-Prefixed String from Memory Address in $B0/$B1
;
; Description:
; This routine outputs a length-prefixed string to the console using the ROM_COUT subroutine.
; The string is stored in memory at the address specified by the zero-page pointers $B0 (low byte) and $B1 (high byte).
; The first byte at this address represents the length 'n' of the string, followed by 'n' characters to be printed.
;
; Inputs:
; - $B0/$B1: Zero-page pointers containing the 16-bit address of the length-prefixed string.
;
; Outputs:
; - The string is printed to the console, character by character, via the ROM_COUT routine.
;
; Registers Used (and modified):
; - A (Accumulator): Used to load string length and characters.
; - X: Used as a counter to keep track of remaining characters to print.
; - Y: Used as an index to traverse the string in memory.
;
; Note:
; - If the calling code requires the values in registers A, X, or Y after this routine executes,
; it should save and restore them before and after the call, respectively.
```As a second data point, Gemini got it wrong, claiming that the string had to be null-terminated, and that the string had to be located at $B0.
Claude (Sonnet) also claimed the string had to be null-terminated, but did see that the string's address was at $B0.
-
💡 Google Pixel 9: presentazione anticipata al 13 agosto al Made by Google
https://gomoot.com/google-pixel-9-presentazione-anticipata-al-13-agosto-al-made-by-google
#AI #android #android15 #blog #eventi #gemini #geminiPro #google #ia #madebygooogle #news #picks #pixel9 #pixel9pro #pixel9proxl #tech #tecnologia #tensorg4 #soc #smartphone #tablet #smartwatch
-
💡 Google Pixel 9: presentazione anticipata al 13 agosto al Made by Google
https://gomoot.com/google-pixel-9-presentazione-anticipata-al-13-agosto-al-made-by-google
#AI #android #android15 #blog #eventi #gemini #geminiPro #google #ia #madebygooogle #news #picks #pixel9 #pixel9pro #pixel9proxl #tech #tecnologia #tensorg4 #soc #smartphone #tablet #smartwatch
-
#AIworm infects users via #AI-enabled #emailclients — #MorrisII #generativeAI worm steals confidential data as it spreads
Named 'Morris II' after original Morris, first computer worm that created a worldwide nuisance online in 1988. This worm targets AI apps and AI-enabled email assistants that generate text and images using models like #GeminiPro, #ChatGPT4.0, and #LLaVA. Researchers successfully tested this Morris II #worm and published its findings using two methods.
https://www.tomshardware.com/tech-industry/artificial-intelligence/ai-worm-infects-users-via-ai-enabled-email-clients-morris-ii-generative-ai-worm-steals-confidential-data-as-it-spreads -
#AIworm infects users via #AI-enabled #emailclients — #MorrisII #generativeAI worm steals confidential data as it spreads
Named 'Morris II' after original Morris, first computer worm that created a worldwide nuisance online in 1988. This worm targets AI apps and AI-enabled email assistants that generate text and images using models like #GeminiPro, #ChatGPT4.0, and #LLaVA. Researchers successfully tested this Morris II #worm and published its findings using two methods.
https://www.tomshardware.com/tech-industry/artificial-intelligence/ai-worm-infects-users-via-ai-enabled-email-clients-morris-ii-generative-ai-worm-steals-confidential-data-as-it-spreads -
#AIworm infects users via #AI-enabled #emailclients — #MorrisII #generativeAI worm steals confidential data as it spreads
Named 'Morris II' after original Morris, first computer worm that created a worldwide nuisance online in 1988. This worm targets AI apps and AI-enabled email assistants that generate text and images using models like #GeminiPro, #ChatGPT4.0, and #LLaVA. Researchers successfully tested this Morris II #worm and published its findings using two methods.
https://www.tomshardware.com/tech-industry/artificial-intelligence/ai-worm-infects-users-via-ai-enabled-email-clients-morris-ii-generative-ai-worm-steals-confidential-data-as-it-spreads -
#AIworm infects users via #AI-enabled #emailclients — #MorrisII #generativeAI worm steals confidential data as it spreads
Named 'Morris II' after original Morris, first computer worm that created a worldwide nuisance online in 1988. This worm targets AI apps and AI-enabled email assistants that generate text and images using models like #GeminiPro, #ChatGPT4.0, and #LLaVA. Researchers successfully tested this Morris II #worm and published its findings using two methods.
https://www.tomshardware.com/tech-industry/artificial-intelligence/ai-worm-infects-users-via-ai-enabled-email-clients-morris-ii-generative-ai-worm-steals-confidential-data-as-it-spreads -
#AIworm infects users via #AI-enabled #emailclients — #MorrisII #generativeAI worm steals confidential data as it spreads
Named 'Morris II' after original Morris, first computer worm that created a worldwide nuisance online in 1988. This worm targets AI apps and AI-enabled email assistants that generate text and images using models like #GeminiPro, #ChatGPT4.0, and #LLaVA. Researchers successfully tested this Morris II #worm and published its findings using two methods.
https://www.tomshardware.com/tech-industry/artificial-intelligence/ai-worm-infects-users-via-ai-enabled-email-clients-morris-ii-generative-ai-worm-steals-confidential-data-as-it-spreads -
Access to Vertex AI Gemini API is currently offered through SDKs not including C#. Visual Studio 2022 comes with an integrated HTTP client using .http files.
https://jochen.kirstaetter.name/getting-started-with-the-vertex-ai-gemini-with-vs-http/
#GeminiSprint #BuildWithGemini #GeminiAI #GeminiPro #gde #MVPBuzz
-
Access to Vertex AI Gemini API is currently offered through SDKs not including C#. Visual Studio 2022 comes with an integrated HTTP client using .http files.
https://jochen.kirstaetter.name/getting-started-with-the-vertex-ai-gemini-with-vs-http/
#GeminiSprint #BuildWithGemini #GeminiAI #GeminiPro #gde #MVPBuzz
-
Getting started with Gemini using Visual Studio (REST API). Access Gemini AI using the integrated REST client and .http files. Works also in VS Code using REST Client extension.
https://jochen.kirstaetter.name/getting-started-with-gemini-using-vs-http/
#GeminiSprint #BuildWithGemini #GeminiAI #GeminiPro #gde #MVPBuzz
-
Getting started with Gemini using Visual Studio (REST API). Access Gemini AI using the integrated REST client and .http files. Works also in VS Code using REST Client extension.
https://jochen.kirstaetter.name/getting-started-with-gemini-using-vs-http/
#GeminiSprint #BuildWithGemini #GeminiAI #GeminiPro #gde #MVPBuzz
-
Google debuts more powerful “Ultra 1.0” AI model in rebranded “Gemini” chatbot - Enlarge (credit: Google)
On Thursday, Google announced that it... - https://arstechnica.com/?p=2001738 #largelanguagemodels #machinelearning #geminiadvanced #textsynthesis #googlegemini #geminiultra #gpt-4-turbo #googlebard #chatgpt-4 #geminipro #microsoft #googleai #ultra1.0 #chatgpt #biz #gemini #google #openai #gpt-4 #bard #ai
-
Google debuts more powerful “Ultra 1.0” AI model in rebranded “Gemini” chatbot - Enlarge (credit: Google)
On Thursday, Google announced that it... - https://arstechnica.com/?p=2001738 #largelanguagemodels #machinelearning #geminiadvanced #textsynthesis #googlegemini #geminiultra #gpt-4-turbo #googlebard #chatgpt-4 #geminipro #microsoft #googleai #ultra1.0 #chatgpt #biz #gemini #google #openai #gpt-4 #bard #ai
-
Google debuts more powerful “Ultra 1.0” AI model in rebranded “Gemini” chatbot - Enlarge (credit: Google)
On Thursday, Google announced that it... - https://arstechnica.com/?p=2001738 #largelanguagemodels #machinelearning #geminiadvanced #textsynthesis #googlegemini #geminiultra #gpt-4-turbo #googlebard #chatgpt-4 #geminipro #microsoft #googleai #ultra1.0 #chatgpt #biz #gemini #google #openai #gpt-4 #bard #ai
-
Google debuts more powerful “Ultra 1.0” AI model in rebranded “Gemini” chatbot - Enlarge (credit: Google)
On Thursday, Google announced that it... - https://arstechnica.com/?p=2001738 #largelanguagemodels #machinelearning #geminiadvanced #textsynthesis #googlegemini #geminiultra #gpt-4-turbo #googlebard #chatgpt-4 #geminipro #microsoft #googleai #ultra1.0 #chatgpt #biz #gemini #google #openai #gpt-4 #bard #ai
-
Google debuts more powerful “Ultra 1.0” AI model in rebranded “Gemini” chatbot - Enlarge (credit: Google)
On Thursday, Google announced that it... - https://arstechnica.com/?p=2001738 #largelanguagemodels #machinelearning #geminiadvanced #textsynthesis #googlegemini #geminiultra #gpt-4-turbo #googlebard #chatgpt-4 #geminipro #microsoft #googleai #ultra1.0 #chatgpt #biz #gemini #google #openai #gpt-4 #bard #ai
-
Google Bard: Here’s how to create your free AI-generated images #ai #digitallife #geminipro #generativeai #google #googlebard #imagefx #imagegenerator #imagen2 #news
https://soyacincau.com/2024/02/02/google-bard-ai-image-generator-imagen-2/
-
Google upgrades Bard with new Gemini AI model, Bard Advanced to arrive in 2024 #bard #bardadvanced #digitallife #gemini #gemininano #geminipro #geminiultra #google #googlebard #googlebardadvanced #googlebardai #googlegemini #googlegemininano #googlegeminipro #googlegeminiultra #news #tech
https://soyacincau.com/2023/12/08/google-upgrades-bard-with-new-gemini-ai-model/
-
Google upgrades Bard with new Gemini AI model, Bard Advanced to arrive in 2024 #bard #bardadvanced #digitallife #gemini #gemininano #geminipro #geminiultra #google #googlebard #googlebardadvanced #googlebardai #googlegemini #googlegemininano #googlegeminipro #googlegeminiultra #news #tech
https://soyacincau.com/2023/12/08/google-upgrades-bard-with-new-gemini-ai-model/
-
Google upgrades Bard with new Gemini AI model, Bard Advanced to arrive in 2024 #bard #bardadvanced #digitallife #gemini #gemininano #geminipro #geminiultra #google #googlebard #googlebardadvanced #googlebardai #googlegemini #googlegemininano #googlegeminipro #googlegeminiultra #news #tech
https://soyacincau.com/2023/12/08/google-upgrades-bard-with-new-gemini-ai-model/
-
Google Gemini vs OpenAI GPT-4 — Bard чи ChatGPT? Демонстрація можливостей нової моделі і кілька наочних прикладів https://itc.ua/ua/novini/google-gemini-vs-openai-gpt-4-bard-vs-chatgpt-agi/ #Генеративнийштучнийінтелект #Сильнийштучнийінтелект #Штучнийінтелект #GoogleGemini #GooglePixel8 #GeminiUltra #GeminiNano #GeminiPro #Україна #ChatGPT #Новини #Gemini #Google #LaMDA #Bard #AGI
-
Gemini Pro in Bard – erstmal wieder nicht in der EU
#Google #BardUpgrade #GeminiPro #KIInnovation #MultimodaleKI #Technologiefortschritt #KünstlicheIntelligenz