Jin Daily AI Trivia: Gemini 2.5 AI is no longer in preview — it’s now stable!
Jin Daily AI Trivia: Gemini 2.5 AI is no longer in preview — it’s now stable!
I used to criticize Google’s Gemini AI for giving overly “DEI-safe” responses (remember the black US presidents incident?). From my experience using it back when it was still called BARD, then Gemini 1 Ultra (which Google claimed could outperform GPT-4 and real humans), the reality was… the responses were unreliable and hallucinations were pretty common.
Fast forward to Gemini 1.5 Flash and 2.0 Flash — they gradually toned down the overfitted DEI replies and the overly diplomatic answers. But honestly, they still weren’t quite on par with other top-tier models like GPT-4.
What really made me subscribing to Gemini Pro is Gemini 2.5 Pro: it has a 1 million token context window and top-tier tool usage, which puts it ahead of many competitors. Plus, Google can lower its costs by running inference on its own TPUs instead of expensive Nvidia GPUs.
Today, Google officially announced that the Gemini 2.5 family is no longer in preview — it’s stable and production-ready now.
So, what does this mean for you?
In short: higher prices. Their cheapest good AI model almost doubled in price. But they also introduced a cheaper option to balance it out.
Here’s the new pricing:
✅ Gemini 2.5 Flash (single pricing now) $0.30 / 1M input tokens (up from $0.15) $2.50 / 1M output tokens (up from $0.60)
✅ Gemini 2.5 Pro $1.25 / 1M input tokens (unchanged) $10.00 / 1M output tokens (unchanged)
✅ Gemini 2.5 Flash-Lite (compared to old Flash) $0.10 / 1M input tokens (down from $0.15) $0.40 / 1M output tokens (down from $0.60)
Comparable models:
🔹 GPT-4o-mini ( vs 2.5 Flash) $0.15 / 1M input tokens $0.60 / 1M output tokens
🔹 GPT-4.1/GPT-o3 (vs 2.5 Pro) $2.00 / 1M input tokens $8.00 / 1M output tokens
🔹 GPT-4.1 Nano (vs 2.5 Flash-Lite) $0.10 / 1M input tokens $0.40 / 1M output tokens
🔹 Qwen-turbo (vs 2.5 Flash-Lite) $0.05 / 1M input tokens $0.20 / 1M output tokens
Google’s feeling the cost squeeze too and is pushing prices up. But luckily, they still have some of the cheapest models compared to OpenAI.
One big plus: Gemini 2.5 Flash-Lite is still the cheapest model that natively accepts audio input — at just $0.50 / 1M input tokens. So if you don’t want to transcribe your audio to text first and need true end-to-end audio handling, Flash-Lite is the best (and almost only) option out there.
hope you learning something today!! See ya!!
