Gemini Omni 1.2 Flash kya hai? Text se 15 second ka AI video kaise banayein (Hinglish guide)

Google ne 5 November 2026 ko Gemini Omni 1.2 Flash ko generally available kar diya. Matlab ab ye test ya preview me nahi, normal use ke liye khula hai. Ye model text aur image se video banata hai, aur video banane ke baad aap usse chat me bolkar edit bhi kar sakte ho. Is post me samjhte hain ki ye kya kar sakta hai, aur ek non-coder kaise pehli baar try kare.

Ye model kya karta hai

Official Gemini API release notes ke hisaab se is model ke main features ye hain:

• Up to 15 second ka video generate karna
• Chat me edit karna, jaise "background badal do" ya "character ko thoda aur slow karo"
• First aur last frame ke beech interpolation (3 se 10 second)
• Video ko aage badhana (extension), per turn 3 se 10 second, total 30 second tak
• Up to 6 reference images se character ya subject ko ek jaisa rakhna
• Output video ke saath native audio, aur 360p se 4K tak resolution

Release notes ke hisaab se video output ki pricing 30% kam hai. Exact rate Google ke pricing page par dekh lo, kyunki rates badalte rehte hain: https://aistudio.google.com/docs/pricing

Pehli baar kaise try karein

Sabse seedha raasta Google AI Studio hai. Model ka code gemini-omni-1.2-flash hai, aur model ka page ye hai: https://aistudio.google.com/docs/models/gemini-omni-flash

Step 1: AI Studio khol kar Gemini Omni 1.2 Flash model chuno. Agar model list mein na dikhe to model page ka "Try in Google AI Studio" button use karo.

Step 2: Ek simple prompt likho. Achha prompt wo hota hai jisme ye teen cheezein ho:

• kaun ya kya dikhna chahiye (subject)
• kahan ho raha hai (setting)
• camera kaise chal raha hai (jaise dheere zoom in, ya side se pan)

Example: "Ek chai wala Mumbai ki local train platform par chai bana raha hai, subah ki halki roshni, camera dheere se zoom in."

Step 3: Pehla result dekho, phir chat mein chhota change bolo. Jaise "ab light thodi warm karo" ya "character ka shirt neela kar do". Ek baar mein sab mat maango, ek ek change se result behtar aata hai.

Step 4: Jo pasand aaye use download karo. Ek generation 15 second tak ka hota hai, extension se total 30 second tak ja sakta hai. Ise short clip ki tarah socho, poori film ki tarah nahi.

Dhyan rakhne wali baatein

• Aap jo video banate ho, uske rights aur usage Google ki terms par depend karte hain. Kisi real insaan ka chehra ya awaaz bina permission ke use mat karo.
• Video mein jo dikhe use real footage bana kar mat pesh karo. Fake footage ko log asli samajh lete hain.
• Model ek naya tool hai. Har baar result pehle se achha aaye, ye zaroori nahi. Do-teen tries lagana normal hai.

Kaun ise use kar sakta hai

Agar aap AI sikh rahe ho, to ye ek achha test hai ki text se video kaise banta hai. Creators ke liye reel ka hook, product ka chhota demo, ya explainer ka ek scene bana sakte ho. Agar aap real estate ya business mein ho, to ek property ka 10 second ka walkthrough concept bhi try kar sakte ho, bas dhyan rahe ki jo dikhe wo asli property jaisa ho.

Ek imaandar note

Ye features Google ke official release notes aur model page se hain. Quality ka faisla aapke apne test se hoga, to ek-do prompts try karke dekh lo.

Source: Gemini API release notes (5 Nov 2026): https://aistudio.google.com/docs/changelog

— Devesh Sarkaar, Founder, Aivesh Academy
Join our WhatsApp channel: https://whatsapp.com/channel/0029Vb8x2347YSd7f4EUue3q