105 Din Baad Gemini Ka Price Double: 3 Cheezein Dekhiye
Gemini 3.8 Flash aaj $0.75 per million input token hai, aur 1 January 2027 se $1.50. Ye Google ke apne pricing page par likha hai. Us din aapke bill ki 3 lines double hongi, aur ek wo hai jise koi dekhta hi nahi.
Agar aap bulk traffic Gemini Flash par bhej rahe hain kyunki wo flagship se kai guna sasta hai, to ye tareekh calendar me daal lijiye. 1 January 2027 ko us model ka har rate double ho jata hai, aur Google ne ye date khud batayi thi jab 2 September 2026 ko Gemini 3.8 Flash aaya.
Farak dekhiye, tabhi baat samajh aayegi. Claude Fable 5.1 aur GPT-6 Astra dono $10 per million input token lete hain. Gemini 3.8 Flash leta hai $0.75 — lagbhag terahwa hissa. Isi liye itna production traffic Flash par baitha hai, aur isi liye double hona pehle se planning maangta hai.
1. Introductory price 31 December ko khatam
Ye koi discount nahi hai jo koi Google ke price ke upar de raha ho. Ye Google ka apna price schedule hai, aur khatam hone ki tareekh launch announcement aur public pricing page dono par likhi hai.
| Rate | 31 Dec 2026 tak | 1 Jan 2027 se |
|---|---|---|
| Input / 1M token | $0.75 | $1.50 |
| Output / 1M token (thinking samet) | $3.75 | $7.50 |
| Cached input / 1M | $0.075 | $0.15 |
| Cache storage / 1M / ghanta | $0.50 | $1.00 |
| Batch input / output | $0.375 / $1.875 | $0.75 / $3.75 |
Gemini 3.6 Flash aur 3.7 Flash bhi usi din double hote hain. Purani generation par jaakar bachne ka raasta kaam nahi karega — ye baat migration plan banane se pehle jaan lena behtar hai, warna mehnat karke bachat zero milegi.
2. Jo line zyadatar log bhool jate hain, wo cache hai
Input aur output sab dekhte hain. Cached read bahut kam log dekhte hain, aur agentic kaam me — jahan assistant har turn par wahi system prompt, wahi documents, wahi badhti hui baat-cheet dobara padhta hai — bill ka sabse bada hissa wahi hota hai.
Tulna ke liye: Claude Fable 5.1 cached context $0.25 per million me padhta hai aur Sonnet 5 $0.20 me. January ke baad bhi Gemini Flash ka $0.15 in sabse sasta rahega — lekin faasla ghat jayega, aur aaj ke numbers par liya gaya routing ka faisla us badlaav ke baad shayad theek na rahe.
3. Wahi per-token price, phir bhi bada bill
Gemini 3.8 Flash bilkul usi rate par aaya jo 3.7 aur 3.6 Flash pehle se le rahe the — teen generation, ek price card. Padhne me ye muft upgrade lagta hai. Per task, hamesha nahi hota.
Har per-token tulna me yahi jaal hai, aur ye dono taraf chalta hai: jo model ek baar me kaam khatam kar de wo us saste model se sasta pad sakta hai jise teen baar chalana pade. Kharcha per token nahi, per poora hua kaam naapiye. Bill wahi number bharta hai.
January se pehle kya kar lena chahiye
- Apne asli workload ka hisaab January wale rate par lagaiye, aaj wale par nahi. Pichhle mahine ka token usage lijiye, Gemini wali lines double kar dijiye, aur dekhiye faisla tab bhi tikta hai ya nahi.
- Dekhiye aapke input ka kitna hissa cached hai. Agar zyada hai, to cache-read aur storage wali line aapke liye headline input price se zyada maayne rakhti hai.
- Apne hi kaam par per completed task naapiye, usi effort setting par jo aap sach me use karte hain. Benchmarks aksar us setting par chalte hain jo aap chala hi nahi rahe.
- Mid-December ka reminder laga lijiye. Daam bina bataye badalte hain, aur samajhdari ye hai ki tab Google ka pricing page khud padh lein, is article ke chaar mahine purane numbers par bharosa na karein.
Is pandrah din me do aur cheezein
Gemini 3.8 Live aur 3.8 Live Extended Thinking 15 September 2026 ko aaye — voice agents ke liye bane live dialogue models, jo beech me toka jana sambhal lete hain aur baat karte-karte bhasha badal sakte hain; Extended Thinking wala version multi-step kaam ke liye hai. Developers ke liye Gemini API aur AI Studio me, aur baaki sabke liye Search Live aur Gemini Live me. Indian audience ke liye bhasha badalne wali baat sabse dilchasp hai — hum ise khud test karke likhenge, launch wali copy dohrayenge nahi.
GPT-6 Astra OpenAI ka pehla model ban gaya jise cybersecurity me Critical level par rakha gaya — unke Preparedness Framework ke tahat, 17 September ko report hua. Usi din Microsoft ne ise Foundry Models me generally available kar diya. Isi classification ki wajah se rollout pehle sirf vetted organisations ko mila tha, sabko nahi.
Save this summary as an image or share it.
AICreatorHub Team
The AICreatorHub editorial team is a group of hands-on AI practitioners, writers and developers based in India. We test AI tools and models ourselves, track official releases from OpenAI, Anthropic, Google, Meta and xAI, and translate them into simple, India-first guides in English and Hindi. Every article is written for real Indian use cases — pricing in rupees, free-tier tips and practical, tested steps — so you get accurate, up-to-date and genuinely useful AI information.