Google's Gemini Pro 2.5 is the best AI coding model right now. I...

@cjzafir
CJ Zafir@cjzafir
44 views Apr 06, 2025 ~3 min read
Advertisement
1
Google's Gemini Pro 2.5 is the best AI coding model right now.

I use Gemini 2.5 + Sonnet 3.5 + GPT o1 models in my new coding workflow.

After using it for 20 hours, here're my finding: ↓
Media image
2
I admit, Google has done a wonderful job this time with Gemini Pro 2.5 model.

They competed with Claude's Sonnet models for coding and the code quality is far better.

In new workflow I use:
- 80% of the time Gemini Pro 2.5
- 10% of the time Sonnet 3.5
- 10% of the time GPT o1
3
My findings after testing Gemini for 20+ hours:

1. Gemini is great for planning docs, understanding large codebases and one-shotting code snippets.

But there're some issues.

It struggles to understand the user's requests sometimes. (You need to be super clear)
Media image
4
2. Gemini's 1M context is not actually 1M context.

If you attach a large codebase (over 500k tokens) it struggles and start hallucinating.

Under 400K tokens it works smoothly.
5
3. Docs are super important.

Gemini model is powerful but it needs deep context about your project.

Attach a knowledge base of coding docs:
- PRD
- App flow doc
- Tech stack doc
- Frontend guidelines
- Backend structure
- Project rules/windsurfrules

This is must! otherwise it starts to hallucinate and mess up the file structure.
Media image
6
4. Gemini Pro works brilliantly with @CodeGuidedev

1. I just generated all my coding docs in 10 minutes.
2. Attached docs in .mdc format (under .cursor/rules)
3. Open Cursor Agent
4. Asked it to follow "implementation plan"

Gemini started coding step by step. It was amazing.
0:38
7
5. Use mixture of models for rare cases.

I found that Gemini can also get stuck in a loop of errors same like Sonnet 3.5/3.7.

To fix this i use GPT o1 or Sonnet 3.5 to scan codebase and just file the 'core issue' and then ask Gemini to implement that fix.

It works.
8
6. Gemini's one shotting is the best.

If you attach your coding docs and provide proper context.

Then Gemini Pro 2.5 model is capable of one-shotting full feature implementation.

I found Cascade works better than Cursor Agent for one shotting scenarios.


@itsPaulAi
Paul Couvert@itsPaulAi
So Gemini 2.5 Pro is now 100% free?!

You can literally use the world's best AI model and one-shot web apps.

All you need is a free Google account:

- Select 2.5 Pro (experimental) from the list
- Activate the Canvas option at the bottom (advanced)
- Send your prompt and enjoy

Example you can reuse:

"Develop a fully functional customer relationship management (CRM) web application tailored for a mid-sized healthcare provider. The application should streamline patient communication, appointment scheduling, record management, and billing processes. As a team of experienced full-stack web developers specializing in healthcare technology solutions, you will design and build this application. The application should feature a user-friendly interface with distinct dashboards for administrators, doctors, and staff. Include functionalities such as patient profiles with medical history, a messaging system, insurance verification, and report generation on key performance indicators like patient acquisition and appointment adherence. The front-end should be built using React. Maintain a professional, solution-oriented, and technically precise tone throughout the development process."
9
7. Games and Illustrations using threejs

Gemini is superiorly better than Sonnet 3.7 when using threejs to code games or any illustrations.

It can one-shot physics concepts and plane games with no sweat.


@rakyll
Jaana Dogan ヤナ ドガン@rakyll
The game is getting together for the first time. Projects, people, meetings, some ping pong... Gemini 2.5 is the first model that took me this far.
10
8. Pricing of Gemini Pro 2.5

Much cheaper than it's direct competitor (Sonnet 3.5/3.7)

Input $1.25 for prompts less than <= 200k tokens
$2.50 for prompts > 200k tokens

Output $10.00, for prompts < 200k tokens
$15.00, prompts > 200k tokens
Media image
11
9. Rate limits

Google launched this model to test the waters and soon they found out that this model is getting so much traction.

So they worked hard (massive respect this time) and increased the rate limits (RPD) for all paid tiers.

Now you see Gemini Pro 2.5 preview model with better rate limits per day.

Plus they made it available for free (with rate limits)
Media image
12
Wrap up:

1. Gemini Pro 2.5 > Sonnet 3.5/3.7 for coding.
2. It can hallucinate. To fix attach coding docs. Use @CodeGuidedev
3. API pricing is alot cheaper.
4. Output token speed is also faster.
5. Code quality and structure is also better.
6. Use it in Cursor & windsurf.
Actions
What You Can Do
  • Export as PDF or Markdown
  • Batch Export to Notion
  • Bookmark & Highlight
  • LinkedIn & Instagram Carousel Maker
Create Free Account

Includes 7-day Premium trial

Advertisement