Ask HN: $50 monthly budget, which coding models would you recommend now?

[dead]

Given the hard budget, I'd throw the $30 on API credits and use those as overflow. If you can do the $100 plan though, that would be optimal. It's a frustrating game though. I'd love to see a 50 dollar plan. At least Claude has a 100 plan, unlike openai.

Claude for the default 5-hour limits. Cursor for the monthly limits when those run out. Cursor is one of the more efficient ones for dumb uses... if you're typing out your own tests, it should work great.

A lot of people recommend OpenAI Codex lately, but I feel it's more suited to vibe coding, where you're giving it a very high level idea.

Hey man, here's how to properly spend your 50 bucks and get the most bang for the $$:

opencode go - $10 a month gets you decent allowance with latest GLM, MiniMax, and Kimi. Windsurf - $20 gives u access to nearlly all models.

For me, I'm having these two subscriptions plus: shared Codex plans & shared Kimi coding Allegretto from Chinese resellers for around $10.

I only use the Claude Opus/Sonnet inside Windsurf after everything is crystal clear, or I want to let it handle tricky stuff where other models can't address. I use GLM/Kimi for everything else. And Windsurf offers free models all the time.

Codex? I know OpenAI is really politically unpopular right now, but it has very high usage limits for the $20 plan. Claude ($20) and Codex ($20) are hard to beat in terms of pure value. Just set Codex on Thinking-High/Extra-High, and it is Opus 4.6 levels for sure (although both have their niche, where they're superior).

Just get a ChatGPT $20/month subscription and use the codex-cli. I’ve been using it every day since November and never run into limits

You can swicth to Sonnet 4.6. It is even an fantastic model for coding, without changing you subscription. Here the 5 hours window limit is much more larger, if using it with specific tasks.

If you want to change the LLM, my advice is Kimi K2 that's much more chip and very skilled for coding.

(you are right: Opus is a very token-consuming even in my experience, but often it is over-skilled and you don't need it)

Get a few GitHub Copilot accounts at 10 bucks each and use OpenAI models, they have very high usage limits I honestly struggle to even hit with one 10 dollar subscription.

What is ur daily coding time? Based on this time I could recommend a good stack that might save ur money as well.

I would keep Claude code for the heavy lifting and then opencode zen (10usd) with Kimi k2.5 for the rest.

[dead]

GitHub Copilot Pro plus gives you access to many models and if you use them a lot, Microsoft loses money (though they can brag about one more active user and get your data).

It doesn’t give you the largest context windows, but most models perform poorly when the context is very big so it’s not a dealbreaker to me.

[flagged]

Setup

Cursor – $20/month Kilocode plugin (inside Cursor) – Free models Google Antigravity Pro Opencode – Free models

Workflow - Backend Development Use Cursor for major backend tasks Flow: Plan → Execute - Frontend Development Use Google Antigravity Pro for major frontend tasks Flow: Plan → Execute - Minor Tasks & Bug Fixes Use: Opencode Kilocode (inside Cursor)

Remote / On-the-Go Setup A Telegram-configured Opencode instance is set up on mobile Used for: Minor tasks Quick fixes while away from the system

Setup 2-3 accounts to use codex. Claude requires at least max plan to avoid hitting limits constantly.

[dead]

[dead]

[dead]

Right now I think GPT 5.4 beats Opus 4.6 and everything else. I get a lot from just the $10 Github Copilot, somehow it hasn't run out yet and i've been going at it insane everyday

windsurf has already switched to the standard billing method. Cursor's self-developed models are quite cost-effective. The most cost-effective one is Codex, the only downside is that it's not good at writing front-end code

[dead]

If it's technical heavy/complex then Codex. It's by far the smartest coding model. You get solid mileage with the $20/month plan

When I'm in the mood to code I'd pay for API request to guarantee I get a response right away. No subscription usage limit is holding me back from making progress. Agents don't work. You have to compose your own context, which means you need to send the raw request. Not have assistant figure out. They are using multiple requests but every request is the end of each conversation. People have no idea how to use these things. That's why there's such gamut of disagreement about how useful this is, because less than 1% of people can benefit from using LLMs, and it's only programmers. So they'll keep increasing prices until only few people willing to pay for it. This is more useful than solidworks, and solidworks is $4000. So this will eventually reach and exceed $10000/year. Only programmers will be able to afford it. Everyone who is now using it for search and entertainment will be out of luck.