MiniMax No Longer Frontier Tier | Generated by AI

Home 2026.07

Question: Is MiniMax no longer a first-tier AI model company/model?

Answer: Mostly yes, if you define “一线模型” as the global frontier tier (GPT-5.x / Claude / Gemini / DeepSeek-V3/R1 level). But saying “彻底不是” is too absolute.

The current landscape is more like:

Frontier tier

These companies are competing at:

MiniMax position

MiniMax was very strong around:

But compared with the current frontier race, it has fallen behind in perception because:

  1. Less breakthrough model releases

    The market moved very quickly:

    GPT-4
       ↓
    Claude 3/4
       ↓
    Gemini 2.x
       ↓
    DeepSeek V3/R1
       ↓
    reasoning + agent era
    

    MiniMax did not create a comparable “everyone talks about this model” moment.

  2. Open-source impact is weaker

    DeepSeek changed the game because researchers and developers could run, fine-tune, and analyze it.

    MiniMax models are less central in the open ecosystem.

  3. Developer mindshare dropped

    For engineers building agents:

    GPT
    Claude
    Gemini
    DeepSeek
    Qwen
    ...
    MiniMax
    

    MiniMax is usually not the first choice.

However, “not frontier” does not mean “bad”.

MiniMax still has:

A more accurate ranking:

Global frontier:
OpenAI / Anthropic / Google / DeepSeek

Strong second tier:
Qwen / Mistral / Meta Llama / MiniMax / Zhipu / Moonshot

Application-focused:
many specialized startups

The interesting thing is that China AI ranking changed a lot after DeepSeek. Before:

Baidu
Alibaba
Tencent
MiniMax
Moonshot
Zhipu

was often discussed.

After DeepSeek:

DeepSeek
Qwen
ByteDance Seed
...

became much more visible.

So your statement is directionally right: MiniMax is no longer considered a top-3/top-5 frontier model player. But it is still a serious AI company, not a failed one.


Back Donate