GosuCoder - Videos
Back to Channel
Top AI Coding Agents Dec 2025 | Opus 4.5, Gemini 3.0 Pro, GPT 5.1
This last month was wild with Gemini 3.0 Pro and Opus 4.5 shaking things up. AI assisted coding has gotten crazy good. Links: 🧑💻My Recommended AI Engineer course is Scrimba: https://scrimba.com/...
AI Coding has changed so much with Gemini 3 Pro and Opus 4.5
I’ve been heads-down coding for 12 to 14 hours a day for the last 9 days straight. It was a grind, but it gave me the perfect opportunity to stress-test the newest AI models in a real-world, high-p...
Cursor 2.0 might have me coming back to Cursor
I really do think this changes the entire landscape for AI coding agents with all the massive price changes combined with them now owning more of their own destiny with less dependency on Anthropic...
From Free to $300 the best options for AI Coding
The best AI coding tools at each tier from FREE to $300 Links: 🧑💻My Recommended AI Engineer course is Scrimba: https://scrimba.com/the-ai-engineer-path-c02v?via=GosuCoder My Links 🔗 👉🏻 Subscrib...
Is Haiku 4.5 good at coding?
Claude Haiku 4.5 came out of nowhere, so after taking some time to use it, and running evals on it. I've come to the conclusion that its a decent model. Links: 🧑💻My Recommended AI Engineer course...
Cursor's plan mode is way better than I expected it to be
I wanted to try a few different plan modes include Droid (Factory CLI), Claude Code, and Cursor. Cursor's newly released plan mode is amazing. Links: 🧑💻My Recommended AI Engineer course is Scrim...
The worst pricing change in AI coding assistants?
As many of you know i've been a fan of Augment Code, highly recommending them over the last year. But this is probably the biggest nerf, especially to legacy customers. I hope they roll back some ...
The most bizarre stealth coding model release so far | Cheetah in Cursor
Stealth AI Coding models are so much fun, i love picking them apart, testing them and trying to figure out who they are. Cheetah is by far one of the most bizarre. Links: 🧑💻My Recommended AI Engi...
OpenAI Dev Day - Agent Builder, Image Editor, Codex updates?
Hoping for some good updates for us coders.
OpenAI Dev Day - Agent Builder, Image Editor, Codex updates?
Hoping for some good updates for us coders.
Best AI coding Agents with some interesting results | Sonnet 4.5, GLM 4.6, Grok Code, GPT 5 Codex
There has been so much in the month of September. I do my best to put them all through the testing gauntlet and share the results here. Fortunately Sonnet 4.5 and GLM 4.6 made it just in time to in...
Qwen just keeps winning, Qwen 3 Max is incredible
Qwen 3 Max puts serious pressure on Anthropic, its an incredible model at coding. Links: 🧑💻My Recommended AI Engineer course is Scrimba: https://scrimba.com/the-ai-engineer-path-c02v?via=GosuCod...
Mystery Coding Model Supernova is made by...
I love stealth models, but i'm pretty confident I know who made this one, and its not Anthropic and its definitely not Sonnet 4.5 Links: 🧑💻My Recommended AI Engineer course is Scrimba: https://s...
GPT 5 - Codex is a fantastic AI model for coding, but...
OpenAI's Codex tuned GPT-5 model is fascinating but its plagued by slowness. Links: 🧑💻My Recommended AI Engineer course is Scrimba: https://scrimba.com/the-ai-engineer-path-c02v?via=GosuCoder ...
Claude Code has some major problems... has it gotten any better?
The exodus of Claude Code or Claude Max in particular has been interesting to watch. While I still subscribe to Claude Max I do have serious concerns. Links: 🧑💻My Recommended AI Engineer course i...
Qwen's biggest AI model Qwen 3 Max... but should you use it to code?
Qwen 3 Max is the biggest model Qwen has released with over 1 Trillion parameters, but can it be your daily coder? Links: 🧑💻My Recommended AI Engineer course is Scrimba: https://scrimba.com/the-...
Qwen 3 Max coding, Stealth Models on OR, hanging out.
Checking out some of the new models while trying to get some work done. Qwen 3 Max Sonoma Models on OR
Kimi K2 got a massive upgrade, possibly the best open source coding model now?
I think Kimi K2 is incredibly good, overtime we'll see how it compares with Qwen 3 Coder. But I showcase what I've learned, and discuss pricing, demos, and prompt caching. Links: 🧑💻My Recommended...
Remote Coding Agents that run locally... super cleaver solution!
Remote Agents have changed a ton in the landscape of coding, and I don't see that slowing down. In a short 3 months from the time I first started covering Remote agents to now, we actually have a g...
Qwen 3 235B on the Framework Desktop for Coding?
I'm super excited with the research I've done with local AI models on the Framework desktop. I showcase what I've learned and discuss the downsides to using Agent based coding with slow LLM models....
Best AI coding Agents with some crazy upsets | GPT 5, Grok Code Fast, Claude, Qwen 3 Coder
There has been so much in the month of August, Grok Code Fast, GPT 5, Kiro, Qoder, Augment CLI and more. I do my best to put them all through the testing gauntlet and share the results here. Some ...
Is GPT-5 and Codex worth using now?
I spent time with the new Codex IDE extensions, tested how it connects to Codex Remote agents, and used the Codex CLI. And I have to say I really enjoyed working with them. GPT-5 seems snappier, th...
OpenAI Dev announcement: Codex update or something else? 10:00 PT
Checking out OpenAI's latest announcement for devs. Could be the Codex updates they rolled out yesterday, or maybe something else surprising.
Can a Local LLM REALLY be your daily coder? Framework Desktop with GLM 4.5 Air and Qwen 3 Coder
With the arrival of my new Framework Desktop I decided to move to coding just with Local LLM's without touching any Claude, GPT5, etc models. I learned a lot while running GLM 4.5 Air, Qwen 3 Coder...
DeepSeek v3.1 Update is Better than I Expected... BUT?
Deepseek v3.1 was an unexpected update, but it comes with some really good good upgrades. First its so much better at tool calling that it actually runs in sst/opencode, it also supports anthropic ...
Is Qwen 3 Coder currently being held back? Most complicated provider testing I’ve done!
I pumped well over 200 million tokens into Qwen 3 Coder provider variance testing. And the results were not what I expected. Links: 🧑💻My Recommended AI Engineer course is Scrimba: https://scrimb...
I think I know who made the latest stealth AI coding model Sonic!
These Stealth AI models are so much fun, I spend so much time thinking about who could of made it. Sonic lives up to its name as its fast, but can it code, and who do I think made it. Links: 🧑💻My...
Are Local LLM's finally good at coding now... Qwen 3 Coder 30b
Local LLM's for Coding have gotten insanely good, Devstral Small 2507, OpenAI's GPT OSS 20B, and Qwen 3 Coder 30b are put head to head using some new tools all running on my RTX 5090 Links: 🧑💻My ...
This open weight model is a coding beast.. GLM 4.5
GLM4.5 is actually a great open weight model, but its held back because of its limited context window. Links: 🧑💻My Recommended AI Engineer course is Scrimba: https://scrimba.com/the-ai-engineer-...
GPT 5 is confusing.... it took me too long to figure this thing out
I used GPT 5 for 50 million tokens, and learned a lot. The biggest lesson is how much faster it is when you configure it at low reasoning and low verbosity. Links: 🧑💻My Recommended AI Engineer co...
The Results are in for GPT 5...
GPT-5 has been hyped beyond belief, lets see how it actually stacks up. Links: 🧑💻My Recommended AI Engineer course is Scrimba: https://scrimba.com/the-ai-engineer-path-c02v?via=GosuCoder My Lin...
GPT 5 at 11:00 AM PT?
OpenAI hopefully is releasing GPT 5, lets find out live as we watch OpenAI's livestream and hopefully get to test it out.
OpenAI FINALLY releases open weight models, but can they actually code?
I spent so much time testing OpenAI's latest open source models, and I just keep learning more and more about the variance in providers. It makes it so hard to really evaluate these things. But I s...
Open AI FINALLY open sources AI models
OpenAI releases two decent sized models, lets throw some stuff at it.
Qwen 3 Coder at 2000 tokens per second and a reasonable price, too good to be true?
I'm torn on Cerebras's offering of Qwen 3 Coder at $50 a month and $200 a month. On one hand I think more AI providers should do this, in fact I've thought OpenRouter should do this, but on the oth...
Is Horizon Beta GPT 5, and Can it Code?
Horizon Beta is one of the most interesting models that i've ever tested. It has just a unique personality. Lets go through some simple tests and how it actually works in real code. Links: 🧑💻My R...
The Best AI Coding Assistants | August 2025... interesting results
Testing 19 different Coding agents including Augment Code, Claude Code, RooCode, and many more along with 5 different AI models: o3, Gemini 2.5 Pro, Claude 4 Sonnet, Qwen 3 Coder, and Kimi K2 this ...
A beautiful terminal experience that can code - Charm Crush
Charm Crush is the latest AI coding assistant CLI, but with a twist its freaking beautiful. The UI experience almost makes it feel like a little mini desktop app. Its missing features and has some ...
Horizon Alpha - is it GPT 5
Another mystery model appears
Horizon Alpha - is it GPT 5
Another mystery model appears
Qwen 3 Coder is almost on par with Claude 4 Sonnet, but...
Frist look at Qwen 3 Coder's evals, and how well it works in Roo Code, Windsurf, Cline etc. But there are issues that need to be known. Links: 🧑💻My Recommended AI Engineer course is Scrimba: htt...
I did not expect Claude Code to release Custom Sub Agents...
Claude Code is having a lot of trouble lately with its outages, and it seemingly taking shortcuts that it wasn't taking before. But it doesn't change the fact that Custom Sub agents is an awesome a...
Qwen 3 Coder is INCREDIBLE but...
Qwen3 Coder is incredibly good at coding, but there are some caveats. It has to be configured right, and it can be costly. Links: 🧑💻My Recommended AI Engineer course is Scrimba: https://scrimba....
AI Coding is changing and not in a way I like!
Amp Code, and Augment Code both don't allow any model selection and I think its a mistake. Lets talk about what this means to us as developers and why we should think of LLM's as programming langua...
AI Writes Garbage Code?
This video is a bit different but I wanted to share my thoughts on how AI Assisted coding impacts each of us. The progress we all go through as we embrace it more, and then to counter the negativit...
Open Weight models are finally getting good at coding...
Mistral really dropped the ball by not open sourcing Devstral Medium when it was released. Kimi K2 has significantly surpassed it in popularity and quality. But Devstral Medium is still a great mod...
Kimi K2 - Open Weight AI actually competes for CODING now!
Kimi K2 open weight AI model is massive, amazing, but slow. Lets go over how well it can code! Links: 🧑💻My Recommended AI Engineer course is Scrimba: https://scrimba.com/the-ai-engineer-path-c02...
Grok 4 is good ... but can it code?
Grok 4 releases as a new top LLM, but what I care about is how does it work with coding, more specifically AI coding assistants. Links: 🧑💻My Recommended AI Engineer course is Scrimba: https://sc...
Grok 4 Release! ... Checking out Perplexity Comet while waiting
Grok 4!!
Top AI Coding Assistants with Gemini 2.5 Pro... Wild variance
Gemini 2.5 Pro is an incredibly hard model to test, as its variance is so high! In this video I test a lot of AI coding assistants like Zed, Cursor, Windsurf with Gemini 2.5 Pro. Links: 🧑💻My Rec...