I had the chance to use both the 8B and 70B version and I noticed the accuracy in responses are like never before. It is far better than it's previous Llama 2 version. It handles document reading and parsing, code generation, debugging and also calculations very well. The speed with which it returns output is also great. Review collected by and hosted on G2.com.

The 8B size is too small, and the 70B is more than necessary which could get a little expensive. They could have released better sizes. The AI model from my experience is way better than Llama 2 but still lacks in reasoning generation when compared to GPT 4o and other similar AI bots. Review collected by and hosted on G2.com.

Thanks to my 4090 the performance is quite well, the variety features and performance is a dignity competitor for OPENAI, I didn't expect this peak, but I continue reading and learning how to handle in the correct way. Review collected by and hosted on G2.com.

Curve learning is pretty high and high consumption, but I think this is normal. These technologies are not easy. Review collected by and hosted on G2.com.

Its ability to handle human-like conversation is absolutely amazing. In terms of communication it is the best chat ai out there that I've used. Review collected by and hosted on G2.com.

It lacks a larger data based knowledge like Gemini for example when running something like a report. Review collected by and hosted on G2.com.

What I like best about Meta Llama 3 is its ability to generate accurate and coherent responses, making it highly effective for various applications. Review collected by and hosted on G2.com.

So far, I have not encountered any issues. Review collected by and hosted on G2.com.

I've found Meta Llama 3 to be incredibly helpful. It can access the latest information from the internet and has a very human-like touch, which helped me make an informed decision about buying a watch. It also assisted me in setting a realistic budget for a future purchase and finding something with the specific features I was looking for.

When I made a big career transition, I used Meta Llama 3 to prepare for interviews and seek guidance on navigating challenging situations. It helped me sound knowledgeable about topics I previously knew nothing about.

I frequently use it to refine and receive feedback on my writing, as well as the feedback I provide to my subordinates. It helps me maintain a fair, empathetic, yet assertive approach.

Overall, I now consider it my go-to platform whenever I have questions or need to share something that's on my mind. I never thought I would rely on AI so confidently! Review collected by and hosted on G2.com.

I have noticed a couple of areas that need improvement. First, the way search results are presented could use some refining. While it's great to see links related to the search, they need to be presented in a way that shows more refined effort and doesn't look like a simple Google search result.

Additionally, it's frustrating when a result disappears while I'm reviewing it, and I receive a message saying "Sorry, I can't help you with this request right now." It becomes repetitive and annoying.

I also think it would be helpful to have the conversation split into groups, making it easier to find past searches without any hassle. For example, if I searched for something two weeks ago, it's currently difficult to locate. Review collected by and hosted on G2.com.

Ease of setup, documentation and community support are best aspects of Llama 3 Review collected by and hosted on G2.com.

Sometimes the contextual accuracy and relevance are not at par which gives unexpected results Review collected by and hosted on G2.com.

Most of the AI image generation tools I've used require you to wait a while before images are actually generated. The Meta AI changes the game quite a bit the way your first image is continually generated on the fly, near instantly, as you type your prompt. Generating additional images still takes a little bit, but the ability to see what you will generally get on the fly is a powerful advantage. Review collected by and hosted on G2.com.

I haven't run into many downsides yet. It's free to use. It does have a soft limit, but it's quite high and does reset after a while. The biggest issue is that the aspect ratios of images generated cannot be changed, but that can be resolved through the use of other AI tools. Review collected by and hosted on G2.com.

When reading or hearing most of the hype surrounding AI, the places I frequent on the web along with folks in my circles all typically spoke to the uses and/or benefits of the competition. Upon giving Meta a whirl though, I was pleasantly surprised. The image generation was swift, and the results were better than some tools created by firms that specialize in imagery. Review collected by and hosted on G2.com.

I did find Meta's ability to remember my preferences, style, and tone to be lacking. Other than that though, it's speedy and fairly accurate. Review collected by and hosted on G2.com.

The best things about it is that they are open source and its chat platform is on par with OpenAI's ChatGPT Review collected by and hosted on G2.com.

They are still not at the level of GPT 4o and lacks multimodality like gemini, claude and GPT Review collected by and hosted on G2.com.

After using it quite frequently, we've found that the 70-B version of the LLM follows instruction very well (better than GPT 3, which has a higher chance of missing instruction) and is very quick to respond considering the size of the model. Grammatically, it performs quite well as well. Review collected by and hosted on G2.com.

Privacy concerns, some countries are not allowed access to the official Facebook platform, although it is technically implementable as an open source LLM.

It also tends to stick to a specific style of writing, which can be a downside for some. Review collected by and hosted on G2.com.