EN

Opus 4.5 Just Destroyed Gemini 3 Pro...

#AI Tools
2025-11-26 19:249 min read

Content Introduction
Ask Questions

In this video, the presenter discusses the release of the Opus 4.5 model, expressing excitement as it stands out among recent releases like Gemini 3 and GVT 5.1. They highlight its superior performance, including achieving a perfect score on the spelt benchmark, showcasing good coding capabilities out of the box. The presenter shares their experience with building and implementing projects using Opus, emphasizing its ease of use for both frontend and backend development. They express satisfaction with its affordability compared to previous models and detail specific tasks accomplished effortlessly using Opus. However, they note some drawbacks, such as issues with the cloud environment and external tools needing setup. Despite these challenges, the presenter overall appreciates Opus 4.5's convenience in increasing productivity and enabling efficient workflows.

Key Information

  • The speaker has recently taken an interest in Opus 4.5, which they believe to be the best AI model currently available.
  • Opus 4.5 received a 100% score on the Spelt benchmark, indicating its strong capabilities in generating code.
  • The speaker has extensively tested Opus 4.5, finding it effective in both front-end and back-end development tasks.
  • The presenter mentions the pricing of Opus 4.5 as reasonable compared to previous models, which were deemed overpriced.
  • The model is noted for its ability to write effective code without needing extensive setup or additional tools like an MCP.
  • The speaker compares the overall performance and user experience of Opus 4.5 with other models like Gemini 3, indicating that Opus feels more efficient and less prone to errors.
  • There are criticisms regarding the state of cloud-based coding tools currently available, especially regarding their lack of LSP support.
  • The speaker recounts their experiences of adapting code and setting up environments using Opus 4.5 and expresses satisfaction with its output fidelity.

Timeline Analysis

Content Keywords

Opus 4.5

Opus 4.5 is described as the best model released recently, distinguishing itself from previous models such as Gemini 3 and GVT 5.1. It has received a perfect score on the spelt benchmark, showcasing its effectiveness in generating code and completing various tasks.

Benchmarking Models

The speaker discusses the limitations of benchmarking models and the necessity for real-world usability rather than solely theoretical performance metrics. They express a preference for models that work well in practical scenarios.

Cloud Code

Cloud Code is mentioned as a platform providing a better user experience compared to others, with the ability to execute tasks efficiently, despite some initial setup challenges and limitations.

Pricing & Subscription Models

The discussion highlights pricing strategies in technology models, particularly noting the affordability of Opus 4.5 compared to its predecessors, advocating for reasonable pricing to encourage widespread usage.

User Experience

User experience with different AI models is central to the discussion, emphasizing that ease of use, intuitiveness, and efficiency are critical aspects that influence the speaker's preference for Opus 4.5 over others.

Programming Tools Integration

The speaker illustrates the integration of various programming tools and features in cloud applications, emphasizing the importance of seamless functionality in enhancing overall productivity.

Limitations of Tools

Identified are some ongoing challenges within programming models, such as the inadequacy of type checking in specific integrated tools, which adds complexity to the coding process and counteracts the benefits of ease of use.

Related questions&answers

What is Opus 4.5?

Opus 4.5 is described as the third best model ever made, which has garnered significant interest and positive feedback in recent discussions.

How does Opus 4.5 compare to previous versions?

Opus 4.5 is seen as definitely the best of the three current models, delivering better performance and functionality.

What are some features of Opus 4.5?

Opus 4.5 performs exceptionally well on the spelt benchmark, scoring 100%, and is capable of generating code more effectively than other recent models.

What improvements does Opus 4.5 offer over Opus 4.1?

Opus 4.5 has a significantly reduced price model, costing a third of what Opus 4.1 used to charge, making it much more accessible for daily usage.

How did the speaker feel about the performance of Opus 4.5?

The speaker expressed satisfaction with Opus 4.5, stating it felt great for day-to-day tasks and offered a more engaged experience than previous models.

What are the pricing details for Opus 4.5?

Opus 4.5 is priced reasonably at 5 in and 25 out, which is a significant reduction compared to former pricing structures.

Is the integration with development tools good?

The speaker mentioned that the integration with existing development tools like Open Code and Cloud Code is functional, though they encountered some initial setup challenges with pricing and limits.

How does Opus 4.5 perform in backend tasks?

Opus 4.5 also managed to successfully handle backend tasks, showcasing its ability to integrate and manage complex code requirements effectively.

What are some challenges faced with other models?

The speaker highlighted that models like Gemini 3 and GBT often perform well in benchmarks but fail to deliver satisfactory day-to-day usability due to quirks and issues in their operation.

What was the speaker's experience with user interface generation in Opus 4.5?

The user interface generation was noted to be effective, with Opus 4.5 able to refactor basic sites into more user-friendly formats without extensive manual intervention.

What does the speaker think about the prospects of cloud models versus traditional coding practices?

The speaker believes cloud models have advantages in usability and integration, but notes they still prefer a good coding setup that allows for clear and easy tuning of the model's behavior.

More video recommendations

Share to: