# MiniMax subscription

73 Hacker News commenters recommended MiniMax subscription and 25 criticized it, November 2025 to November 2026. It was recommended in 13 of 13 months. It meets the Hacker Approved badge rule.

**Verdict:** Recommended as a cheap, strong alternative to Claude and other frontier models, especially for coding and agentic work. Common caveat: slow outputs and excess reasoning tokens.

## Why people recommend it

- Low cost for heavy use (11 comments): "I now make it a point of subscribing (directly) with providers that are transparent like MiniMax, DeepSeek, Xiaomi, & Z.ai." ([source](https://news.ycombinator.com/item?id=48961565)) "After 75% price cuts by DeepSeek & Xiaomi MiMo (and now MiniMax), I pretty much packed my Claude bag up and moved over." ([source](https://news.ycombinator.com/item?id=48623365)) "budget alternatives from the families of DeepSeek, Kimi, GPT, MiniMax, etc. which are very much Sonnet grade" ([source](https://news.ycombinator.com/item?id=48545240))
- Near frontier quality for many tasks (7 comments): "The current SotA for open models (dsv3, glm, minimax, devstral, etc) are at or above the mini versions of top labs" ([source](https://news.ycombinator.com/item?id=46734923)) "Many of them (K2.5 the Minimax, the GLM models, the Qwen 3.6 models) are about as capable as frontier US models" ([source](https://news.ycombinator.com/item?id=47741965)) "There are larger open weights models that are more capable, such as minimax and glm5.1." ([source](https://news.ycombinator.com/item?id=48225567))
- Good for coding and agentic work (6 comments): "Kimi, Mimimax and GLM models provide far more robust coding assistance at sometimes no cost" ([source](https://news.ycombinator.com/item?id=47765586)) "MiniMAX and Claude models tend to be better AI models at doing agentic tasks." ([source](https://news.ycombinator.com/item?id=47751163)) "Minimax has been great for super high speed web/js/ts related work." ([source](https://news.ycombinator.com/item?id=46548082))
- Fast output speed (2 comments): "Minimax has been great for super high speed web/js/ts related work." ([source](https://news.ycombinator.com/item?id=46548082)) "I didn't notice slower outputs, it's even faster than Anthropic." ([source](https://news.ycombinator.com/item?id=47540508))

## What people warn about

- Slow output and heavy reasoning (5 comments): "To this day Minimax and GLM are very strong coders but have the same issue, inference speed." ([source](https://news.ycombinator.com/item?id=50156175)) "Sitting around 15 minutes for local Minimax is stupid when you're trying to be productive." ([source](https://news.ycombinator.com/item?id=49243041)) "The down sides emerge pretty fast: much higher reasoning token use, slower outputs" ([source](https://news.ycombinator.com/item?id=47538089))
- Falls short on complex tasks (5 comments): "In my harness Minimax wasn't even getting half of them right, whereas Sonnet was 100%." ([source](https://news.ycombinator.com/item?id=47677371)) "for anything even remotely complex these models seriously struggle." ([source](https://news.ycombinator.com/item?id=47630352)) "they've all been barely half as capable as a middling Codex model" ([source](https://news.ycombinator.com/item?id=47654929))
- Gets stuck in tool-call loops (2 comments): "MiniMax is kind of horrible that it somewhat regularly fall into loops" ([source](https://news.ycombinator.com/item?id=48328280)) "MiniMax got stuck with looping through broken tool calls all the time" ([source](https://news.ycombinator.com/item?id=48291922))

## Alternatives hackers recommend

- [Claude Opus](https://hackerapproved.com/p/claude-opus.md): 3331 recommend
- [OpenAI GPT](https://hackerapproved.com/p/gpt-5-5.md): 1205 recommend
- [Claude Sonnet](https://hackerapproved.com/p/claude-sonnet.md): 854 recommend
- [DeepSeek V4 Flash](https://hackerapproved.com/p/deepseek-v4-flash.md): 828 recommend
- [Fable](https://hackerapproved.com/p/fable.md): 823 recommend

## Where to get it

- [Visit the official site](https://www.minimax.io/)

## Cite as

Hacker Approved, "MiniMax subscription", https://hackerapproved.com/p/minimax/ (counts from 1,795,640 Hacker News comments, November 2025 to November 2026).