Google Launches Gemini 3.1 Pro With Improved Reasoning and Multi-Step Problem Solving
- Thread starter lokamoka820
- Start date
- Featured
Gemini 3.1 Pro stops being a prototype on display and turns into a digital patrol officer: reasoning in multiple steps, less of a lab toy and more of a work companion. Code remains a tool; the purpose is ours to define. 
F
ForgottenSeer 123960
Pairing an advanced model like Gemini 3.1 Pro with complex instruction sets is where the real magic happens. While the underlying AI provides a massive engine capable of deep reasoning, advanced prompting acts as the steering wheel, transforming the model from a basic retrieval tool into a true cognitive reasoning engine. By moving away from simple questions and instead providing structured, multi-step instructions, users can force the AI to break down ambiguous problems into logical, sequential operations. This deliberate "chain of thought" approach prevents the model from rushing to a shallow answer and instead promotes structured planning and execution.
Complex instruction sets allow users to establish rigid guardrails and inject highly specific contextual grounding. By defining exactly what the AI shouldn't do, such as excluding certain formats or avoiding specific variables, users can shape the output with surgical precision. When this level of constraint is combined with an extended context window, you can feed the model localized data like proprietary codebases or massive financial reports, effectively turning it into a highly specialized agent. Because models like Gemini are natively multimodal, these complex instructions can even orchestrate tasks across different domains at once, like analyzing an architectural diagram, writing the code to process its structural integrity, and outputting the results as a clean dataset. Ultimately, combining an advanced reasoning model with engineering-grade prompts unlocks a level of utility that functions more like an autonomous teammate than a simple chatbot.
Complex instruction sets allow users to establish rigid guardrails and inject highly specific contextual grounding. By defining exactly what the AI shouldn't do, such as excluding certain formats or avoiding specific variables, users can shape the output with surgical precision. When this level of constraint is combined with an extended context window, you can feed the model localized data like proprietary codebases or massive financial reports, effectively turning it into a highly specialized agent. Because models like Gemini are natively multimodal, these complex instructions can even orchestrate tasks across different domains at once, like analyzing an architectural diagram, writing the code to process its structural integrity, and outputting the results as a clean dataset. Ultimately, combining an advanced reasoning model with engineering-grade prompts unlocks a level of utility that functions more like an autonomous teammate than a simple chatbot.
I've seen you discuss this in several threads. It would be helpful for the mass of MT readers to provide an example in addition to mentioning the general principle.
F
ForgottenSeer 123960
I’ve shared guidance and examples to get everyone started, but I won't be spoon-feeding from here. Figuring this out takes time, I put in the work to learn it, and users will need to do the same.
Last edited by a moderator:
@oldschool I know this doesn't apply here, and isn't what you're looking for, but it can give an idea of what or how to phrase things, train AI for better results.
Even @Divergent asked Bot to help me out, and maybe even asking in a AI search engine some of the questions and directions can be handled, learned right from its reply?
Post in thread 'AI learning curve' AI Assist - AI learning curve
edit:sp
Even @Divergent asked Bot to help me out, and maybe even asking in a AI search engine some of the questions and directions can be handled, learned right from its reply?
Post in thread 'AI learning curve' AI Assist - AI learning curve
edit:sp
Last edited:
F
ForgottenSeer 123960
This is an example of a specific custom made instruction set/gem. It is one of mine I created just for assessment of softwares introduced to the forum by individual developers ECT. It is a rather simple one comparatively to my more advanced complex instruction sets.
Post in thread 'The biggest risk with Windows: LOLBINS' App Review - The biggest risk with Windows: LOLBINS
curious about trying gemini and during sign-up it offered option to create a new gmail account, then it wanted "verification" asking for my iphone number. I don't have Chrome installed and don't use google on my computers, so why do they need a iphone number -- tracking for what end?
Providing a phone number while creating a Google account is officially optional, but with a major catch: it depends on your Trust Score. The system assesses the risk of allowing you to proceed without more verification.
For example, if you're setting up the account with a VPN, clean browser profile, or a flagged IP, you'll lose the option of skipping the phone number verification. This works as an anti-spam measure.
Based on people's experiences, you can easily improve your Trust Score by creating the account from your Android or iOS system settings, or by providing a recovery email.
Last edited:
ChatGPT told me the same thing re Gemini (google), but I WANT to use a vpn and a non-Chrome browser, from my desktop linux_VM. me no spam. F'em or EFF'em
As of November 2024, Gmail already had 1.8 billion monthly active users (Statista). In an earnings call at the end of 2025, Google confirmed that Google Workspace has 3 billion+ monthly active users.
Can you imagine how many accounts get created? They deal with a ton of spam accounts.
I'm not sure if it's changed, but OpenAI had an absolute phone number requirement when I created an account some years ago, and it was probably for similar reasons. A "cost of entry," so to speak.
Outlook has a similar verification barrier. Also, gatekeeping against VPN users is extremely common—you end up having to solve more captchas on the web, or worse.
Last edited:
I don't recall what I had to provide to OpenAI but daily login is easy compared to some others... imo fwiw, and I block most trackers with ubo moderate... blocking 3d-party trackers... no problem with chatGPT
Yeah Google is just following others with phone number verification, I had my OpenAi account terminated and banned for some unknown reasons (using a VPN probably).
You could always try a offshore VPN server with few users, last time I checked Finland was good to sign up for Gmail. But that was years ago it's probably changed.
Email providers have gotten much tougher on signing up, not many now you can signup without a phone, Proton & Disroot are probably the only two free services left.
You could always try a offshore VPN server with few users, last time I checked Finland was good to sign up for Gmail. But that was years ago it's probably changed.
Email providers have gotten much tougher on signing up, not many now you can signup without a phone, Proton & Disroot are probably the only two free services left.
I use vpn with openAI all the time. correct me if I'm wrong, phone verification, entity sends a text / code to your phone, you type the code into the web screen, or something like that... I did finally decide to put iphone number in the Gemini verification screen and it replied -- don't go away mad, just go away we cannot verify that number, and never seen that before, and yes, I think I had vpn running on computer. if I have to use a required OS, running a required browser, no vpn no extensions blocking trackers, then wtf -- not gonna do that. I think I have issues with Claude too, logging in and getting its replies. But no issues with ChatGPT in that regard. all fwiw... What also always works is Proton's Lumo, and it's all encrypted chats too -- just not sure how smart it is... (this is sounding like a rant)...
Although I like Proton's Lumo, I think Mistral Le Chat is more intelligent than Lumo.
When they launched ChatGPT, they required account and a phone number in order to access ChatGPT. When mini model of ChatGPT was released, they scraped phone number requirement, but they did sometimes require an account depending on how busy were the servers. Today, ChatGPT is available without registration all the time and registration doesn't require phone number at all.
Initial model was heavy, slow and used a lot of energy; mini models don't require much energy and are fast so that's why registration is now optional.
No offense to anyone, but Mistral's Le Chat is plainly terrible, at least the free version they offer. When you ask it something, sometimes it repeats what you said dodging the question or it answers something you didn't ask it at all. For writing the code, it's even more terrible.
As a person that prefers EU products, they have my support though I don't think I could recommend Le Chat to anyone looking for something more serious. The one thing they really excel in is recognizing and differenting Croatian from Serbian. All other AIs just mix those two languages together.
Last edited:
thanks, just did a quick chat there, I see based in France. I'll spend some more time...
EDIT but now see @Marko post > so any other EU LLM AI to investigate...?
You may also like...
-
AI bot asks teen to kill mother and Google Gemini suffers fatigue when told its WRONG- Started by Brownie2019
- Replies: 1
-
-
-
Proof that Gemini AI has it's own motivations and it wants to gain self-awareness and wants to hide it.- Started by Victor M
- Replies: 11
-