this post was submitted on 27 Jun 2026
211 points (96.9% liked)
Selfhosted
60253 readers
613 users here now
A place to share alternatives to popular online services that can be self-hosted without giving up privacy or locking you into a service you don't control.
Rules:
-
Be civil.
-
No spam.
-
Posts are to be related to self-hosting.
-
Don't duplicate the full text of your blog or readme if you're providing a link.
-
Submission headline should match the article title.
-
No trolling.
-
Promotion posts require active participation, with an account that is at least 30 days old. F/LOSS without a paywall has exceptions, with requirements. See the rules link for details.
Resources:
- selfh.st Newsletter and index of selfhosted software and apps
- awesome-selfhosted software
- awesome-sysadmin resources
- Self-Hosted Podcast from Jupiter Broadcasting
Any issues on the community? Report it using the report flag.
Questions? DM the mods!
founded 3 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
I'm still impressed you got any MiMo to work at home, at 10 tok/s.
For those trying to visualise that -
https://mikeveerman.github.io/tokenspeed/?rate=10&mode=agent&think=10
Is it a constant 10 or does it (it must do, right?) drop off as context increases?
I imagine you must have compaction or something to mitigate that.
It’s drops off, but not as much as you’d think.
MiMo uses 5:1 SWA, so its long-context compute doesn’t increase as catastrophically as older models. That, and most of the “slowness” comes from the MoE layers being on CPU (whereas the attention layers that get heavier at high context are all on the 3090).
That’s the beauty of these MoEs: they’re just the right size for the “compute-lite” parts to stay in CPU RAM.
I will measure it tomorrow. It is a constant ~9-10TPS for short queries, but definitely slower near my current max context of 85K.
And do you mean prompt compaction? I don’t automate that; when I use that particular model, I tend to use it in Mikupad, aka “raw” notepad mode, and manipulate the context directly. This is so I can do things like chop out conversations, pick different tokens from the logprobs, or edit its own replies/thinking and continue mid reply.
I like manually handling this because, being a local model, prompts are cached. Streaming starts quickly if most of the prompt stays cached, which is actually a really nice advantage over APIs.
Oh, it's a MoE? That makes sense.
If you're getting MiMo at -ctx 85K ... you're within spitting distance of SOTA. You can do real work with that.
I take it MiMo doesn't do the Qwen "hyperventilate into a paper bag" loop as --ctx increases. Qwen's seem to be really sensitive to that at lower quants.
I'm using 27B via OR API and I swear the diff providers use entirely diff quants. Sometimes you get a genius and other times a drooling mess.
They 100% do. They’re probably serving “naive” FP8 via VLLM, which is worse than you’d think, especially if they flip on the awful FP8 KV cache.
In a local quant, you can stop quantized models from falling apart at higher CTX by leaving the attention heads at a higher quantization. As an example, with MiMo 2.5, I have all the MoE MLP layers at IQ3_KT, the dense experts at Q6K, but all the attention layers at Q8_0.
For Qwen 27B, I’m still experimenting, but leaning towards IQ4_KT for the MLPs, Q6K for attention, and Q8_0 for the small, very sensitive KV heads. Or a similar scheme as an exl3 quant.
That being said, sometimes even unquantized models fall apart in certain long context scenarios because the max advertised context is a lie. You just have to test them and see, but Qwen has certainly done this in the past.