DeepSeek, new AI model from China, it's the new rival to the US AIs. It's a game changer: you can sign up today and it's open sourced! It's free.
- Thread starter World B Free
- Start date
Ask it practically anything.
Yesterday, my friend came by to help me with one of my AC units. It stopped blowing cold like a week ago. My friend, is older, learned the trade last year has all the tools and stuff, then got hurt last year and stopped doing it. He gets help from time to time from a master HVAC guy, but just to not constantly call the master tech guy every little question, we asked Grok what the SLT and LLT readings meant on his AC Manifold Gauge. It explained it. Then said it would get a better understanding if I gave it the high and low side readings, I did. It said those are not right, told me what they should be. Asked me the outside temperature...
Then it spit out 3 options of why it failed. Said it's either the TXV (Thermal Expansion Valve) or the filter drier or low refrigerant.
We know the TXV was good and so were the refrigerant levels. So it dialed down that the filter drier was clogged. Told us the answer with just diagnosing from a few prompts.
Then even went on the tell us, shut off the unit, wait for pressures to equalize from high and low sides, we did. Then said if you turn it back on once equal and you see the high side shoot up relatively quickly, then your filter drier is clogged.
That chit blew me away yesterday. That was the issue...
Yes, IF you know what you're doing and ONLY if you know what you're doing, you can use Claude to help you build software. Cause 50% of the time, it will be wrong and you will have to correct it.
Another great Chinese AI
www.z.ai
www.z.ai
- tartup Z.ai, formerly known as Zhipu, announced Monday that its new GLM-4.5 AI model would cost less than DeepSeek to use.
- Like DeepSeek, the new model is also open source and can be downloaded for free.
- At about half the size of DeepSeek’s model, GLM-4.5 only needs eight Nvidia H20 chips to operate, Z.ai CEO Zhang Peng told CNBC on Monday.
Can someone curate a list of ai that’s out?
Honestly, I’m only using ChatGPT and Claude.
I primarily use Claude, it’s been great for me. But I’m open to explore others, especially for ADVANCED coding.
Honestly, I’m only using ChatGPT and Claude.
I primarily use Claude, it’s been great for me. But I’m open to explore others, especially for ADVANCED coding.
| ChatGPT | GPT-3.5, GPT-4, GPT-4o | Freemium | General purpose, content creation, coding, problem-solving.cnet.com |
| Google Gemini | Gemini 1.5 Pro, Gemini 1.5 Flash | Freemium | Information queries, document analysis, content generation, coding.cnet.com |
| Claude | Claude 3 series (Opus, Sonnet, Haiku), Claude 3.5 Sonnet | Freemium | Conversational AI, long-context processing, coding, privacy-focused tasks.cnet.com+1 |
| Perplexity AI | Proprietary models, GPT-4, Claude 3 | Freemium | AI-powered search engine, information retrieval, and summarization.cnet.com |
| Microsoft Copilot | OpenAI's GPT-4, GPT-4o, and other models | Freemium | Integrated AI assistant in Windows and Microsoft 365, web search, content creation.cnet.com |
| Jasper | GPT-3.5, GPT-4, Anthropic's Claude, proprietary models | Paid | Marketing, sales, and business content creation with brand voice.softonic.com |
| You.com | GPT-4, Claude 3 series, Gemini Pro, Zephyr | Freemium | AI-powered search with customizable AI model selection.toolsforhumans.ai |
| ChatSpot | GPT-3, GPT-4 | Free | AI assistant for HubSpot customers, sales and marketing tasks.softonic.com |
| Intercom | GPT-4, proprietary models | Paid | Customer service automation, lead qualification, and user engagement.softonic.com |
| Freshchat | Freddy AI (proprietary), Microsoft Azure OpenAI Service | Paid | Customer self-service, automating support conversations.softonic.com |
| Crisp Chatbot | Proprietary LLM | Paid | Lead nurturing and customer messaging for businesses.softonic.com |
| Kommunicate | GPT-4 | Paid | E-commerce customer support and automation.softonic.com |
| Chatsonic by Writesonic | GPT-4, GPT-4o mini, Claude, Gemini, Stable Diffusion | Freemium | Real-time data and trend-based content and image creation.wheelhouse.com+2 |
| HuggingChat | Open-source models (e.g., Llama) | Free | Open-source alternative for conversational AI and various tasks.larksuite.com+1 |
| Pi | Inflection-2.5 (proprietary) | Free | Empathetic and supportive personal AI for conversations and advice.techradar.com+2 |
| Socratic by Google | Google AI | Free | Homework help and learning for students.softonic.com+2 |
| Character.AI | Proprietary neural language models | Freemium | Entertainment, role-playing, and creating customizable AI characters.cnet.com+2 |
| GitHub Copilot | GPT-4, GPT-4o, Claude 3.5 Sonnet, Gemini 1.5 Pro | Paid | AI code completion and programming assistance for developers.team-gpt.com+1 |
| Amazon CodeWhisperer | Proprietary LLMs trained on Amazon and open-source code | Freemium | Real-time code suggestions and security scans for developers.aihungry.com+2 |
| Tavus API | Proprietary models | Paid | AI-powered personalized video generation for marketing and sales.tavus.io+2 |
| Quidget | Not specified | Paid | All-around performance for accuracy, ease of use, and integrations. |
| Botpress | Various LLMs (integrates with major providers) | Freemium | Developer-focused platform for building customizable chatbots.botpress.com+2 |
| Rasa | CALM (Conversational AI with Language Models) | Freemium (Open Source) | Pro-code framework for building custom, enterprise-grade AI assistants.rasa.com+2 |
| Meya | GPT-3.5-turbo | Paid | Platform to build and deploy ChatGPT-powered bots for apps and websites.siteefy.com+2 |
| Pandorabots | Proprietary | Freemium | Platform for building and deploying chatbots with API integrations. |
| DeepSeek | DeepSeek-V3, DeepSeek-R1 | Freemium | Open-source models for general-purpose chat, reasoning, and coding.larksuite.com+2 |
| xAI - Grok | Grok 4, Grok 4 Heavy | Paid | Real-time information access via X (formerly Twitter), complex reasoning.api.chat+2 |
| Mistral - Le Chat | Mistral Large, Mistral Small | Freemium | High-performance, efficient AI assistant with a focus on privacy.techradar.com+2 |
| Cohere | Command R, Command R+, and other proprietary models | Paid (with a free tier for developers) | Enterprise-focused AI for chat, text generation, and search.cnet.com+2 |
| Alibaba - Qwen | Tongyi Qianwen (Qwen) series | Open Source | Multilingual large language models for various AI applications.toolsforhumans.ai+2 |
| Reka AI - Yasa | Reka Spark, Reka Flash, Reka Core (proprietary) | Paid | Multimodal AI assistant for enterprise applications. |
[th]
[th]
[th]
[th]
Chatbot
[/th][th]
Language Model(s)
[/th][th]
Free/Paid
[/th][th]
Primary Use
[/th]
Paid
| Jasper | February 2021 | GPT-3.5, GPT-4, Anthropic's Claude, proprietary models | Marketing, sales, and business content creation with brand voice. |
| Intercom | 2011 (AI features like Fin released in 2023) | GPT-4, proprietary models | Customer service automation, lead qualification, and user engagement. |
| Freshchat | 2017 (Freddy AI introduced later) | Freddy AI (proprietary), Microsoft Azure OpenAI Service | Customer self-service, automating support conversations. |
| Crisp Chatbot | 2015 | Proprietary LLM | Lead nurturing and customer messaging for businesses. |
| Kommunicate | 2017 | GPT-4 | E-commerce customer support and automation. |
| GitHub Copilot | June 2022 | GPT-4, GPT-4o, Claude 3.5 Sonnet, Gemini 1.5 Pro | AI code completion and programming assistance for developers. |
| Tavus API | 2021 | Proprietary models | AI-powered personalized video generation for marketing and sales. |
| Quidget | 2023 | Not specified | All-around performance for accuracy, ease of use, and integrations. |
| Meya | 2015 | GPT-3.5-turbo | Platform to build and deploy ChatGPT-powered bots for apps and websites. |
| xAI - Grok | November 2023 | Grok 4, Grok 4 Heavy | Real-time information access via X (formerly Twitter), complex reasoning. |
| Cohere | 2021 | Command R, Command R+, and other proprietary models | Enterprise-focused AI for chat, text generation, and search. |
| Reka AI - Yasa | April 2024 | Reka Spark, Reka Flash, Reka Core (proprietary) | Multimodal AI assistant for enterprise applications. |
| Drift | 2015 | Proprietary AI | Conversational marketing and sales, lead generation. |
| Zendesk AI | 2023 (Advanced AI add-ons) | Proprietary models, integrates with OpenAI | Customer service automation, agent assistance, and ticketing. |
| LivePerson | 1995 (Conversational AI developed over time) | Proprietary Conversational AI | Enterprise-scale customer engagement and commerce. |
| Ada | 2016 | Proprietary AI | Automated customer experience (ACX) for enterprises. |
| Tidio | 2013 | Proprietary AI (Lyro) | Customer service and sales automation for small to medium businesses. |
| Copy.ai | 2020 | Proprietary models, integrates with major LLMs | Sales, marketing, and workflow automation content creation. |
| Writesonic | 2021 | GPT-4, GPT-4o mini, Claude, Gemini, Stable Diffusion | Content creation for marketing, SEO, and social media. |
| Anyword | 2013 | Proprietary AI, integrates with major LLMs | Performance-driven marketing content generation and optimization. |
| Gong | 2015 | Proprietary AI | Revenue intelligence, analyzing sales conversations and data. |
| Kore.ai | 2013 | Proprietary NLP and LLMs | Enterprise conversational AI platform for various industries. |
| Verloop.io | 2016 | Proprietary NLP | Conversational AI for customer support, marketing, and sales. |
| Haptik | 2013 | Proprietary AI, integrates with Generative AI models | Customer support and engagement for large enterprises. |
| SnatchBot | 2015 | Proprietary NLP, integrates with major LLMs | Platform for building chatbots for various channels. |
| CloserCopy | 2020 | Proprietary AI | Long-form content creation, especially for marketing and sales copy. |
| Outgrow | 2015 | Integrates with GPT models | Interactive content like quizzes, calculators, and chatbots for marketing. |
| MobileMonkey | 2017 | Proprietary AI, integrates with OpenAI | Sales and marketing automation for Instagram, Facebook, and SMS. |
| ManyChat | 2015 | Proprietary AI | Chat marketing automation across various messaging platforms. |
| Salesforce Einstein | 2016 | Proprietary AI models | AI integrated across the Salesforce platform for sales, service, and marketing. |
[th]
[th]
[th]
[th]
Chatbot
[/th][th]
Public Release Date
[/th][th]
Language Model(s)
[/th][th]
Primary Use
[/th]
Hey, awesome! Thanks for the quick response. Imma check some of these out.
From my personal experience, and from what I have read.> Claude is your best bet for coding. They have been steadily implementing less and less daily and weekly limits now (i am on the 20 plan) because of abuse by some persons.
I have been able to do some AMAZING things with claude with very minimal errors or friction. THere is a CLaude Reddit group that drops gems daily.
Initally, I was trying with Chat GPT, and while the conversation windows are longer, it continuously fucked up and made mistakes stupid mistakes that frustrated the hell out of me. So, I use both in tandem. Chat GPT for general queries and strategies etc (no coding) and claude STRICTLY for coding (no small talk)
With all that said, I think you have the best strategy already in place
Keep your eye on Kimi K2 and Qwen3. They may not be better than Claude for coding at the moment, but they could be one version away from taking the lead.
Yeah, Claude has been example in so many ways. ChatGPT is too pushy and generic for me at times, and once you cross a threshold of advance coding, ChatGPT is inferior.
I just submitted a manuscript with the help of both ChatGPT and Claude’s guidance. But the level of editing, research, doubling-checking, and rewriting is a reminder that these are tools and the responsibility is still on the user.
These AI tools don’t know when to leave something alone even if it works perfectly. Lol
At this point, I’m looking for AI that can hold down 2000+ lines of code flawlessly.
Claude is excellent up to about ~800 lines then it begins to fall off if the task is too complex. Ive got a lot better at managing how much code it produces or the task I ask for.
I’ll keep a look out on your suggestions. Thanks, fam!
This is true. These tools are not for the stupid, they are for people who already know what they are doing.
Are hallucinations a problem when using AI for coding purposes? This is a general question based on all of the AI apps you've used. Personally, I've cooled on AI once it started to be confidently wrong about fact-based information.
I don’t know when you mean by hallucinations for AI coding. But Claude does well for me, but I also plan well on what I need.
By hallucinations I meant AI giving false information as if it's true. It's a pretty big problem with AI. I was wondering if that were also the case with using AI for coding.
Yes that is a big problem.
I just listened to a podcast from NPR, they reported that ppl familiar with coding often find it very time intensive to locate the actual line/lines of code that cause the issues and typically report that rewriting the code over from scratch is a better use of their time.
An early 2024 survey found 60% of respondents found ai useful in coding
In an early 2025 survey, that number dropeed to 30%....
It was a fairly recent pod and not that long, maybe 7 minutes....if I have time to look, I'll post it
Actually, no. AI (Claude) is much better about coding than providing information based on facts.
Logic coding is straightforward. AI mostly codes based on what you give it. However, the responsibility is still on the user to double-check their work.
I have developed some deep code with some over 1000 lines without issue.
That’s wild because I have the exact opposite experience.
I used AI coding for Bash, Python, CSS, JavaScript, HTML, PHP, and C++.
Programmers shouldn’t have to get AI to rewrite entire code unless the drafted artifact is mangled, which happens from time to time. However, I have used surgical approaches successfully without having to refactor an entire piece of code.
This is very problematic if people are just deploying code and not knowing what it does fully. lol.
Remember this conversation in a few years when you're cursing your robot out for continuing to put too much salt on your food....
Sidenote: I find it find it so funny when people cuss out these personal assistants...My ex curses out Alexa all the time and I just find it hilarious. "Bitch, I told your dumbass to turn on the lights.."
If you ask Web based LLMs about information that is after their knowledge cutoff, you can force it to search for the most recent information by telling it "make sure your answer is accurate as of August 2025" or something similar. That usually gets them to search for the most recent information.
Claude will usually have you start a new chat if the context gets too long. I don't think I've ever seen it hallucinate. It's not perfect, but it's way better than Gemini. Gemini will be super confident and wrong as hell.
Gemini also has the tendency to get stuck in a loop. It'll give you a wrong answer, acknowledge that the answer was wrong when you point it out, and then give you the exact same answer again.
Remember this conversation in a few years when you're cursing your robot out for continuing to put too much salt on your food....
Sidenote: I find it find it so funny when people cuss out these personal assistants...My ex curses out Alexa all the time and I just find it hilarious. "Bitch, I told your dumbass to turn on the lights.."
Man, can you imagine the verbal lashing these AI assistants catch from people?
Aww man, I can't remember where I saw it, but there was this chick that was cursing out google maps for giving her wrong directions and i was like...maaaaan this is so funny. Of course it's us doing it...Man, can you imagine the verbal lashing these AI assistants catch from people?
Claude be bullshitting sometimes. This mofo told me I solved the "Desert thermal accumulation paradox."
I'm like, WTF is that shit? Of course, it was nothing, but that mofo was confident as hell like I made some breakthrough.
Similar threads
