What is Mellum2?
JetBrains' open 12B mixture-of-experts model for fast routing, RAG and sub-agent steps in coding systems.
Who is Mellum2 best for?
Mellum2 is best for teams building AI developer tools or agent pipelines who need a small, fast model they can self-host.
Routing between models
It reads incoming prompts quickly and picks the right model or tool, so large models only handle the hard requests.
Low-latency RAG
It summarises retrieved context and answers fast because only 2.5B of its 12B parameters are active per token.
Agent sub-steps
Context gathering, planning and validation steps can run on Mellum2 instead of a single large model.
Private deployments
Open weights and vLLM support let you keep code and data on your own hardware.
How much does Mellum2 cost?
Mellum2 is free and open source under the Apache 2.0 license, with weights on Hugging Face to run locally or on your own servers.
- Base, Instruct and Thinking versions
- 131K-token context
- Runs with vLLM; GGUF builds available
JetBrains shows no paid plan for Mellum2: the product page sends you to the Hugging Face weights to run locally or on your own cloud. Full pricing on huggingface.co. We checked these prices on .
How do readers rate Mellum2?
No one has rated Mellum2 yet. Be the first.
Sign in to rate. One rating per account.
Is Mellum2 open source?
Yes. Mellum2's code is public under the Apache 2.0 licence, so you can run it on your own server.
What are the best Mellum2 alternatives?
Mellum2 is an open source alternative to Gemini 3.5 Flash-Lite and GPT-5.4 mini & nano. DeepSeek V4.1 Flash, 1-bit Bonsai, Holo3.1 and Nemotron 3.5 Lightning do the same job and are open source too.
Google's low cost Gemini 3.5 model for high volume tasks and subagents.
OpenAI's small GPT-5.4 models with a 400K context and computer use.
DeepSeek's small, efficient open model.
PrismML's 1-bit language models in 8B, 4B and 1.7B sizes that run on phones and laptops.
H Company's open computer-use models for web, desktop and mobile agents.
NVIDIA's open 30B model with 3B active parameters, built for fast agent execution.
Questions about Mellum2
What is Mellum2?
JetBrains' open 12B mixture-of-experts model for fast routing, RAG and sub-agent steps in coding systems.
Who is Mellum2 for?
Mellum2 is best for teams building AI developer tools or agent pipelines who need a small, fast model they can self-host.
Is Mellum2 free?
Yes. Mellum2 has a free Open weights plan: base, Instruct and Thinking versions.
Is Mellum2 open source?
Yes. Mellum2's code is public under the Apache 2.0 licence, so you can run it on your own server.
What are the best Mellum2 alternatives?
The closest alternatives on Recaply are Gemini 3.5 Flash-Lite, GPT-5.4 mini & nano, DeepSeek V4.1 Flash, 1-bit Bonsai, Holo3.1 and Nemotron 3.5 Lightning.
Where can I try Mellum2?
On its official website, jetbrains.com.
Should I use Mellum2 Instruct or Mellum2 Thinking?
JetBrains suggests Mellum2 Instruct for direct, low-latency answers such as chat, code help and tool calls. Mellum2 Thinking reasons step by step before answering, which suits harder jobs such as complex debugging or planning.







