Create

Sign in to ReadmeX

or

How we handle your data: Privacy Policy

ReadmeX

Your AI reading companion

What would you like to catch up on?

Pick a briefing, or ask your own question.

Your community and people briefings stay personal to you.

Story

Community comparison of seven compact language models

AI summary

A community post compares seven compact language models, from 1.92B to 9B parameters, across specifications, context lengths, modalities, licenses and benchmark results. The models are K2-Horizon-7B, Granite 4.2-8B, Qwen3.5-9B, MiMo-V2.6-Distill-Qwen-9B, LFM2.5-8B-A1B, Gemma 4 E4B and LFM3-VL-2B. The author notes that the scores are lab-reported and that results drawn from different evaluation harnesses should not be compared directly.

In-depth insight

The post consolidates specifications and benchmark claims for several compact models, providing a convenient starting point for comparison while requiring caution because the reported scores come from laboratories and use different evaluation harnesses.

Scores are lab-reported so take them with a grain of salt.

Specs

ModelOrgParams (total/active)ContextModalityLicense
K2-Horizon-7BIFM7B dense512KtextApache-2.0
Granite 4.2-8BIBM8B dense (thinking)128K (512K ext.)textApache-2.0
Qwen3.5-9BQwen9B hybrid (delta-net, thinking)262K (~1M ext.)text + vision + videoApache-2.0
MiMo-V2.6-Distill-Qwen-9BXiaomi9B (SFT of Qwen3.5-9B)~262K (base)text + vision (visual coding)MIT
LFM2.5-8B-A1BLiquid8.3B / 1.5B active MoE128KtextLFM 1.0
Gemma 4 E4BGoogle4.5B effective / 8B total (PLE)128Ktext + image + audioApache-2.0
LFM3-VL-2BLiquid1.92B (1.73B backbone, thinking)128Ktext + image + videoLFM 1.0

Text benchmarks

Superscripts: a = K2 card (IFM harness), b = Liquid harness, c = MiMo card (Xiaomi harness); no mark = own card; e = from LFM3-1.7B, the text backbone; d/† = MiMo avg@3 / internal set. Bold = best in row (cross-harness caveat applies).

Some source content is omitted. Continue reading on the original site.

#K2-Horizon-7B#Granite 4.2-8B#Qwen3.5-9B#MiMo-V2.6-Distill-Qwen-9B#LFM2.5-8B-A1B#Gemma 4 E4B#LFM3-VL-2B#IFM#IBM#Xiaomi#Liquid#Google

Original · 2 min readRead the original
Group chat

No comments yet. Start the conversation.