GPT 4o vs. Sonnet 3.5 for Coding?

[Marco69]

Must ask you guys, can free chatgpt coding / create a tool embedd it on my wp then it works for everyone who visit my site?
Like a keyword tool or an indexer, etc,etc,etc....?

 

[Next Press Pro]

Although I find Claude to be generally superior to ChatGPT I did find it randomly rejecting lots of perfectly benign prompts recently. This is something Anthropic needs to fix ASAP.

 

[arcticsocks]

I use gpt for coding and everything suits me

 

[adarshak103]

What are the prerequisites to run a local LLM on our computers?

What level of python do I have to know ?

checkout ollama, easy to setup

 

[Next Press Pro]

I do find ChatGPT to be better at writing -- even compared to Sonnet 3.5.
But I still think Claude is significantly better at coding and critical thinking.

 

[Roger Marquez]

Have you try the new LLAMA 3.1 405B parameters locally? How about using GPT4ALL? What PC specs would you need?

 

[Next Press Pro]

Have you try the new LLAMA 3.1 405B parameters locally? How about using GPT4ALL? What PC specs would you need?

Nope, I've never tried it.
I've been sticking to commercial LLMs such as ChatGPT and Claude (Opus & Sonnet 3.5)
Although I think it would be an interesting development to monitor as Meta states "Our experimental evaluation suggests that our flagship model is competitive with leading foundation models across a range of tasks, including GTP-4, GPT-4o, and Claude 3.5 Sonnet..""

 

[Tolonfel]

Programmer here! Sonnet sounds great for coding, your Opus experience is convincing. Definitely worth a try!

 

[Vaginigi]

Both OpenAI and Anthropic are doing amazing things, exciting competition! Sonnet 3.5 might be a real competitor to the mysterious GPT-5.

AI has changed my coding life: brainstorm ideas, write boring code, even catch bugs! It's a powerful tool, not a replacement for coders though.

 

[harrysharp]

Neither. Run a local LLM specifically trained for programming like DeepSeek-Coder-V2, Codestral, or CodeLlama. That'll get you better results than either, especially if you can run the larger models (above 7b) that are less prone to hallucinating. You'll need a lot of VRAM though (or you got to wait longer with a CPU). The poor mans way is pickup an M40 or P40 and swap the cooler from a GTX 770 then you can run it in a regular desktop. 24GB VRAM lets you offload most LLM's to the GPU, with only the largest hitting system RAM too.

OpenAI vs Anthropic? OpenAI is going to win, they're better at marketing. It doesn't matter which product is better, the best product often doesn't win. Most people don't know anything but ChatGPT even exists.

Anyone claiming Llama is better than Sonnet for coding better come with receipts. That's a lot to back up... though I'd like to see it.

 

[SirSmash]

Anyone claiming Llama is better than Sonnet for coding better come with receipts. That's a lot to back up... though I'd like to see it.

I have nothing to prove just because you come here and demand receipts. If you don't agree then you don't agree. Doesn't make one lick of difference to my life.

 

[WAndreas]

GPT 4o v. Sonnet 3.5

I tried a lot of things before settling on GPT4o. For me there is no alternative to it.

 

[k9pilot44]

Step 1. Go to https://openrouter.ai/
Step 2. Create an account, deposit $10 (crypto or CC) copy your API key
Step 3. Go to Lobechat.com
Step 4. Create an account, pay for cheapest tier ($4/mo iirc)
Step 5. Go to “Language Model” in settings, scroll to open router section, paste api key, add the models you want (o1, o1-mini, Claude 3.5 Sonnet, Chat GPT 4o, llama, mistral etc)
Step 6. Just chat, create assistants, or import assistants on the market (many coding assistants out there, search GitHub for system prompts)

 

[xDyablo]

For all the BHW programmers out there..

What's your take on GPT 4o v. Sonnet 3.5 for programming task?

Personally, I have found the code produced by Claude Opus 3.0 to be superior than ChatGPT. After several weeks of failing to resolve a programming issue with ChatGPT, I was able to solve it in less than one hour with Opus 3.0. Honestly, I'm blown away by Anthropic.

What's your take on OpenAI vs. Anthropic? Do you think Claude Opus 3.5 will rival GPT-5?

How has AI changed your coding workflow?

Sonnet 3.5 all the way , i use them both , but claude sonnet 3.5 far superior than Gpt4o

 

[Nikolaos]

For all the BHW programmers out there..

What's your take on GPT 4o v. Sonnet 3.5 for programming task?

Personally, I have found the code produced by Claude Opus 3.0 to be superior than ChatGPT. After several weeks of failing to resolve a programming issue with ChatGPT, I was able to solve it in less than one hour with Opus 3.0. Honestly, I'm blown away by Anthropic.

What's your take on OpenAI vs. Anthropic? Do you think Claude Opus 3.5 will rival GPT-5?

How has AI changed your coding workflow?

I prefer Sonnet 3.5 the most for coding.

 

[Jerichomiles]

Sonnet 3.5 is way ahead in coding, that's for sure. Not least since ChatGPT seems to keep getting dumber and dumber. He has ridiculous memory problems where three messages down he already doesn't remember what you said or even he said himself. Grok is providing competition though. I just started testing it since yesterday and his coding is possibly on a par with Claude.​

Sonnet 3.5 all the way , i use them both , but claude sonnet 3.5 far superior than Gpt4o

Yeah definitely, I didn't think anyone was even questioning that at this point since it's not even close.

Both OpenAI and Anthropic are doing amazing things, exciting competition! Sonnet 3.5 might be a real competitor to the mysterious GPT-5.

Yeah competition for GPT-5 for sure, GPT4 is certainly way behind Sonnet 3.5. I use AI for coding every day so I have tested them rigorously. Currently testing Grok and am impressed so far but it's only been one day. I think the whole GPT5/AGI thing is just hype to keep you using dumber by the day GPT4.

 

[harry_se9]

Been using both for the past few months, here's my 2 cents...

@SirSmash makes a solid point about local LLMs, but for most day-to-day coding tasks, I'm sticking with cloud solutions.

I've found Claude 3.5 Sonnet better for:

• Complex debugging (way better at understanding stack traces)
• Writing tests
• Code refactoring

GPT-4 shines at:

• Quick snippets and solutions
• React/TypeScript (like @rafark mentioned)
• Frontend stuff generally

The real game-changer isn't which one's "better" - it's using them for different things. I use GPT-4 for quick fixes and Sonnet when I need deeper code analysis or explanation.

Pro tip: Try having both generate the same solution and compare. You'd be surprised how often they complement each other.

That's just my workflow tho, YMMV

 

[Jerichomiles]

Been using both for the past few months, here's my 2 cents...

@SirSmash makes a solid point about local LLMs, but for most day-to-day coding tasks, I'm sticking with cloud solutions.

I've found Claude 3.5 Sonnet better for:

• Complex debugging (way better at understanding stack traces)
• Writing tests
• Code refactoring

GPT-4 shines at:

• Quick snippets and solutions
• React/TypeScript (like @rafark mentioned)
• Frontend stuff generally

The real game-changer isn't which one's "better" - it's using them for different things. I use GPT-4 for quick fixes and Sonnet when I need deeper code analysis or explanation.

Pro tip: Try having both generate the same solution and compare. You'd be surprised how often they complement each other.

That's just my workflow tho, YMMV

Lol you mean ChatGPT is better at easy stuff. In reality Claude can do all the easy stuff too, you're probably just using ChatGPT instead because Claude has limited use. I'll do the same I must admit, I'll leave Claude to do the difficult stuff. That does actually mean Claude is better though, by some distance. At least for coding anyway.

 

[agofufop2015]

Sonnet is better for coding . In my experience may be the best