MiniMax
AI company offering MiniMax large language models via API, coding subscriptions and open weights.
HN commenters like MiniMax for very cheap token plans and pay-as-you-go use1234 and as a capable model for agents, subagents and utility tasks like code exploration.5678 The most common caveat is that it trails Claude and GLM on demanding coding work,910 and its large models are hard to self-host on consumer hardware.111213
Why people recommend it
Low cost · 3
Token plans and pay-as-you-go use are seen as cheap for heavy token use.
“a MiniMax subscription for high-token-use tasks that don’t need the best model (for $20 I get what feels like infinite tokens)”
“Currently testing M3 for agentic tasks. It works OK and their token plan is very cheap. Highly recommend for claw / hermes type of work.”
Show 1 more
“I just got MiniMax $200/year token plan. Usually it works fine for daily coding”
Agentic tasks · 3
Used for subagents, agents and utility roles like exploration and library tasks.
“My current go-tos are DS Flash, Minimax M3 for subagents.”
“I also use MiniMax-M3 in utility roles like explore/library tasks.”
Show 1 more
“MiniMAX and Claude models tend to be better AI models at doing agentic tasks.”
Output quality · 3
Users report surprisingly good results, with less nonsense than some rivals.
“I have been using MiniMax M3 and I am really surprised by the quality this model spits out.”
“When M3 came out, I started using it, and it was actually really good.”
Show 1 more
“I used Claude Code with GLM 5.1, MiniMax M2.7, Kimi K2.6 and had pretty good results.”
Generous limits · 2
The plan supports API workflows and rate-limits instead of suspending heavy users.
“I just got MiniMax $200/year token plan. Usually it works fine for daily coding”
“a MiniMax subscription for high-token-use tasks that don’t need the best model (for $20 I get what feels like infinite tokens)”
Fluent writing · 2
One user enjoyed talking with it for its grounded world view and fluent write-ups.
“I really really enjoyed talking with MiniMax M3 a lot.”
Transparent provider · 2
Recommended as a transparent provider for direct coding subscriptions.
“subscribing (directly [0]) with providers that are transparent like MiniMax, DeepSeek, Xiaomi, & Z.ai.”
“Z.Ai subscription/ Kimi subscription / MiniMax subscriptions are good”
Caveats and dissent
Coding quality · 2
Some find it far behind Claude, ignoring parts of prompts, and a step down from GLM-5.2.
“the quality is really far from Claude models.”
“Minimax M3 feels okay but definitely a step down from GLM-5.2.”
Self-hosting · 3
The models are large, so local runs are slow, need aggressive quantization, or don't fit.
“I suspected that MiniMax M2.5 might be a bit too much for this board.”
“i used to mix remote and local minimax 2.7(q3) on my strix halo”
Show 1 more
“it felt like such a breath of fresh air when I used the reference full-size model”
Context costs · 1
Pay-per-token use gets pricey as the context fills up.
“It's quite cheap and seems to be better than Gemini 3 Flash”
Often compared with
1 moved from Claude Code to MiniMax
1 moved from DeepSeek to MiniMaxGLM10 comments mention both
2 moved from Claude to MiniMax
1 moved from Kimi to MiniMax
Who recommends it in their own words
Threads where it came up
- MiniMax M3.1 Flash 2 mentions
- DeepSeek costs OpenCode Go user $1.14/day; dual DGX breaks even in 24 years 2 mentions
- Kimi K3.1 Pro and K3.1 Pro Highspeed 1 mention
- Claude Haiku 5.5 1 mention
- Ask HN: What default model do you use and why? 1 mention
- Anthropic's best AI model struggles to attract users as cheaper tools thrive 1 mention
Also written as MiniMax M3, MiniMax M2.7, Minimax M3, MiniMax M2.5. Data: minimax.json. Quotes are verbatim; the verdict and theme labels are written by a model from these comments, and every clause links to its sources.
All comments 31
“then handed off to DeepSeek or Minimax”
“I really really enjoyed talking with MiniMax M3 a lot.”
“And M2.7. it had a good world view”
“I also have a $10 minimax plan and use M3 for exploration and library roles”
“My current go-tos are DS Flash, Minimax M3 for subagents.”
“a MiniMax subscription for high-token-use tasks that don’t need the best model (for $20 I get what feels like infinite tokens)”
“I have been using MiniMax M3 and I am really surprised by the quality this model spits out.”
“I've been very surprised with MiniMax M3, not as a particularly good architect, but at it's very good ability to state the problem elegantly”
“stepping up to minimax m3 if that doesn't cut it”
“MiniMax M3 for agents”
“switched to Minimax M3 continuing from the same context”
“Minimax M3 feels okay but definitely a step down from GLM-5.2.”
“subscribing (directly [0]) with providers that are transparent like MiniMax, DeepSeek, Xiaomi, & Z.ai.”
“I configured it to use MiniMax M3 for this task”
“Z.Ai subscription/ Kimi subscription / MiniMax subscriptions are good”
“When M3 came out, I started using it, and it was actually really good.”
“I also use MiniMax-M3 in utility roles like explore/library tasks.”
“it felt like such a breath of fresh air when I used the reference full-size model”
“Currently testing M3 for agentic tasks. It works OK and their token plan is very cheap. Highly recommend for claw / hermes type of work.”
“i used to mix remote and local minimax 2.7(q3) on my strix halo”
“pay as you go on cheaper models like MiniMax you can get a lot done for not a lot of coin”
“I used Claude Code with GLM 5.1, MiniMax M2.7, Kimi K2.6 and had pretty good results.”
“I use Claude Code with GLM, Kimi and MiniMax models.”
“MiniMAX and Claude models tend to be better AI models at doing agentic tasks.”
“It's quite cheap and seems to be better than Gemini 3 Flash”
“the quality is really far from Claude models.”
“I just got MiniMax $200/year token plan. Usually it works fine for daily coding”
“I think only Claude Sonnet/Opus, GPT 5.2+, Minimax M2.5 are useful.”
“Current best models for that - MinimaxM2.5, Gemini 3.1 Flash”
“I suspected that MiniMax M2.5 might be a bit too much for this board.”
“I do believe if there is something that comes up that is literally continuous, would be interesting, but I'm not sure about it right now.”