Gabriel Botelho

desenvolvimento

[Gabriel Botelho @Gabriel53788166]

@gabrielbuzziv Vc já era hypado kkkkkk

[uiuxjosh♦️♦️♦️ @uiuxjosh]

The interface is the product. Treat it that way.

[Gabriel Botelho @Gabriel53788166]

🚨 O Gemini 4 Hydrogen já estreou no TOP 5 do Agent Arena!

O ranking avalia o desempenho de IAs em tarefas reais, considerando sucesso, satisfação, capacidade de correção e uso de ferramentas.

Com +14,8% acima da média de referência, superou o GPT-6 Astra e o Gemini 4 Argon. 👀

[Arena.ai @arena]

Gemini 4 Hydrogen (High) by @GoogleDeepMind debuts #5 on Agent Arena with +14.8% net improvement across 6.5k real-world agentic sessions.

Hydrogen stands out for its speed and cost efficiency:
- Fastest model in the top 10: 5.3 min/task (median), 3x faster than #1 Claude Fable 5.2 and 2.4x faster than GPT-6 Astra
- Cost-efficient: $1.6 cost/task (median), 3.7x cheaper than GPT-6 Astra and 4.7x cheaper than Claude Fable 5.1. API token pricing: $1.5/$7.5 input/output per million tokens
Notably, it spends more output tokens per task: 41% more than Carbon, 2.8x more than Argon.

Strong across key signals:
- #2 Steerability: +21% (behind only Fable 5.2 at +33%; ahead of Carbon at +19% and Opus 5.5 at +18%)
- #4 Praise vs. Complaint: +32%
- #5 Confirmed Success: +15%

Ranking by task category:
- #3 in Chat: +19%
- #7 in Work: +13%
- #9 in Code: +12%

Hydrogen delivers competitive performance while completing tasks significantly faster and at a fraction of the cost of several leading models. Congrats to the @GoogleDeepMind team on another impressive release!

[Gabriel Botelho @Gabriel53788166]

@yukii737 @pablomarcal pescando

[Abolfazl @mrjs6781]

این Tab Bar توی Dribbble ۴۰۰ هزار بازدید داره.

ویدیوی انیمیشنش رو دادم به Claude، یه نسخه‌ی اینتراکتیو گرفتم. فنر، motion blur و قطره‌ی مایع برای سوییچ بین گزینه‌ها.

دمو: https://t.co/3DFpx59jCg
کد: https://t.co/7C4RFMA8qi
طراح اصلی: @Volorf

[Gabriel Botelho @Gabriel53788166]

@hadukemv é dentro, compensa demais.

[Gabriel Botelho @Gabriel53788166]

@Mark_2077 @guieenrico

[Gabriel Botelho @Gabriel53788166]

@Google agora só falta um reset pra galera! kkkkkkk

@Google

[Gabriel Botelho @Gabriel53788166]

Bora de teste ja!

[Gabriel Botelho @Gabriel53788166]

a economia do 3.8 e a eficiencia do 4... sensacional!

obrigado @Google

[Science girl @sciencegirl]

Accurate😂

[Google for Developers @googledevs]

Introducing Gemini 4 Hydrogen

While Gemini 4 Carbon and Argon are built for deep, long horizon reasoning, Gemini 4 Hydrogen brings cutting-edge intelligence to everyone — with speed and efficiency for everyday coding and agentic workflows.

Surpassing Argon on the Artificial Analysis Intelligence Index at its default thinking level, at 2.5x the speed of Gemini 3.8 Flash, Hydrogen is engineered for high-velocity everyday work with:
— Multi-model orchestration & parallel agent workflows in Google @Antigravity
— Four tunable thinking levels for reasoning on demand (low to xhigh)
— Reliable harness execution for iterative coding & debugging
— Multimodal & embodied reasoning at scale

Start building with Gemini 4 Hydrogen today: with the Gemini API in @GoogleAIStudio and @AndroidStudio, client-side via @Firebase AI Logic, in Google @Antigravity for paid Google AI Pro and Ultra subscribers, and in Gemini for business.

Get the details: https://t.co/HJjLsBNA8M

[Gabriel Botelho @Gabriel53788166]

5/5 ⚡ Contexto de 1 milhão de tokens. Na API: US$ 1,50 por milhão na entrada e US$ 7,50 na saída.

Está chegando ao Gemini, Busca e AI Studio. No plano gratuito do Gemini, o uso é ativado automaticamente conforme a tarefa.

Fonte: Google.

[Gabriel Botelho @Gabriel53788166]

1/5 🚨 O Google acaba de lançar o Gemini 4 Hydrogen!

A proposta: entregar IA potente com foco em velocidade, programação e agentes autônomos.

E o mais curioso? Ele já supera o Gemini 4 Argon em um dos principais índices de inteligência. 🧵

[Gabriel Botelho @Gabriel53788166]

4/5 📊 Outros benchmarks do Gemini 4 Hydrogen:

• SWE-sweep (programação): 13,5%
• HLE-Diamond (raciocínio): 58,8%
• MKA (multimodal): 53,4%
• Agents’ Last Exam: 71,6%
Mas ele não vence tudo: no Terminal-Bench 4.1, o Claude Sonnet 5.5 lidera com 63,9% contra 57%.

[Gabriel Botelho @Gabriel53788166]

@gabrielbuzziv @guieenrico

[Utsav Techie @utsavtechie]

Liquid Glass before Liquid Glass