Mellum2.1 12B A2.5B Thinking — Hardware Requirements & GPU Compatibility
ChatFunctions
Mellum2.1 Thinking is JetBrains' mixture-of-experts reasoning model with 12 billion total and 2.5 billion active parameters, built for agentic work in code repositories as well as hard coding, math and reasoning problems. It has 28 layers and 64 experts with 8 active per token, and it is post-trained from Mellum2-12B-A2.5B-Base mostly with reinforcement learning, including training inside real repositories with a shell and file-editing tools. The card reports self-evaluated scores of 82.0 on LiveCodeBench v6, 47.0 on SWE-bench Verified and 83.3 on AIME 25/26. Only 2.5 billion parameters are active per token, so it should decode quickly, and at 4-bit it fits a single consumer GPU or a laptop with enough memory for the full 12 billion weights. The context length is 131,072 tokens, with sliding-window attention on three of every four layers. It is released under the Apache 2.0 license, permitting unrestricted commercial and research use. Published in September 2026, it is the successor to Mellum2 Thinking, with the same architecture and an improved post-training stage.
198 downloads
81 likesSpecifications
- Publisher
- JetBrains
- Family
- Mellum
- Parameters
- 12.1B
- Architecture
- MellumForCausalLM
- Context Length
- 131,072 tokens
- Vocabulary Size
- 98,304
- Release Date
- 2026-09-20
- License
- Apache 2.0
Get Started
HuggingFace
Run in cloud
Fits on RTX 3060 12GB (4 GB headroom) · Q4_K_M
- Generation speed
- ~123 tok/s
- generation speed
- Cost per 1M output tokens
- $0.14
- per 1M output tokens
How Much VRAM Does Mellum2.1 12B A2.5B Thinking Need?
Select a quantization to see compatible GPUs below.
| Quantization | Bits | VRAM | + Context | File Size | Quality |
|---|---|---|---|---|---|
| Q2_Kest. | 3.40 | 5.5 GB | 6.5 GB | 5.16 GB | 2-bit quantization with K-quant improvements |
| Q3_K_Mest. | 3.90 | 6.3 GB | 7.3 GB | 5.92 GB | 3-bit medium quantization |
| Q4_K_M | 4.80 | 7.7 GB | 8.7 GB | 7.29 GB | 4-bit medium quantization — most popular sweet spot |
| Q5_K_Mest. | 5.70 | 9.0 GB | 10.0 GB | 8.66 GB | 5-bit medium quantization — good quality/size tradeoff |
| Q6_K | 6.60 | 10.4 GB | 11.4 GB | 10.02 GB | 6-bit quantization, very good quality |
| Q8_0 | 8.00 | 12.5 GB | 13.5 GB | 12.15 GB | 8-bit quantization, near-lossless |
| BF16 | 16.00 | 24.7 GB | 25.7 GB | 24.30 GB | Brain floating point 16 — preferred for training |
est.= calculated VRAM estimate; no published GGUF file found for that quantization yet. Other rows are verified against real community uploads.
Which GPUs Can Run Mellum2.1 12B A2.5B Thinking?
Q4_K_M · 7.7 GBMellum2.1 12B A2.5B Thinking (Q4_K_M) requires 7.7 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 10+ GB is recommended. Using the full 131K context window can add up to 1.0 GB, bringing total usage to 8.7 GB. 52 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti, NVIDIA GeForce RTX 3080.
Runs great
— Plenty of headroom
Decent
— Enough VRAM, may be tightWhich Devices Can Run Mellum2.1 12B A2.5B Thinking?
Q4_K_M · 7.7 GB50 devices with unified memory can run Mellum2.1 12B A2.5B Thinking, including NVIDIA DGX H100, NVIDIA DGX A100 640GB, iPad Pro M5 13" (16 GB).
Runs great
— Plenty of headroomDecent
— Enough memory, may be tightWhere to Download Mellum2.1 12B A2.5B Thinking
Community quantizations of this model — GGUF for llama.cpp, Ollama, and LM Studio, plus AWQ/MLX variants where available.
Related Models
Frequently Asked Questions
- How much VRAM does Mellum2.1 12B A2.5B Thinking need?
Mellum2.1 12B A2.5B Thinking requires 7.7 GB of VRAM at Q4_K_M, or 24.7 GB at BF16. Full 131K context adds up to 1.0 GB (8.7 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 12.1B × 4.8 bits ÷ 8 = 7.3 GB
KV Cache + Overhead ≈ 0.4 GB (at 2K context + ~0.3 GB framework)
Fit ratings and hardware model lists check this model with room for a 16K-token context, which needs a little more memory.
KV Cache + Overhead ≈ 1.4 GB (at full 131K context)
VRAM usage by quantization
Q4_K_M7.7 GBQ4_K_M + full context8.7 GB- Can NVIDIA GeForce RTX 4090 run Mellum2.1 12B A2.5B Thinking?
Yes, at Q8_0 (12.5 GB) or lower. Higher quantizations like BF16 (24.7 GB) exceed the NVIDIA GeForce RTX 4090's 24 GB.
- What's the best quantization for Mellum2.1 12B A2.5B Thinking?
For Mellum2.1 12B A2.5B Thinking, Q4_K_M (7.7 GB) offers the best balance of quality and VRAM usage. Q5_K_M (9.0 GB) provides better quality if you have the VRAM. The smallest option is Q2_K at 5.5 GB.
VRAM requirement by quantization
Q2_K5.5 GBQ4_K_M ★7.7 GBQ5_K_M9.0 GBQ6_K10.4 GBQ8_012.5 GBBF1624.7 GB★ Recommended — best balance of quality and VRAM usage.
- Can I run Mellum2.1 12B A2.5B Thinking on a Mac?
Yes, but only at lower quantizations. The smallest Mac that can run Mellum2.1 12B A2.5B Thinking is Apple iPhone 17 Pro (12 GB) at Q2_K; 33 of the 39 Macs we list can run it at some quantization. For Q4_K_M (7.7 GB) you need a Mac with more unified memory.
- Can I run Mellum2.1 12B A2.5B Thinking locally?
Yes — Mellum2.1 12B A2.5B Thinking can run locally on consumer hardware. At Q4_K_M quantization it needs 7.7 GB of VRAM. Popular tools include Ollama, LM Studio, and llama.cpp.
- How fast is Mellum2.1 12B A2.5B Thinking?
At Q4_K_M, Mellum2.1 12B A2.5B Thinking can reach ~169 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~257 tok/s. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = 1000 ÷ (active GB ÷ (bandwidth GB/s × efficiency) × 1000 + layers × routing ms)
Mixture-of-Experts: only the active experts are read per token, plus a fixed per-layer routing cost.
Example: NVIDIA B200 → 1.5 GB active ÷ (8000 × 0.65) = 0.296 ms, plus 28 layers × 0.055 ms = 1.540 ms, so 1000 ÷ 1.836 ms = ~545 tok/s
Estimated speed at Q4_K_M (7.7 GB)
~545 tok/s~257 tok/s~545 tok/s~494 tok/sReal-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
- What's the download size of Mellum2.1 12B A2.5B Thinking?
At Q4_K_M, the download is about 7.29 GB. The full-precision BF16 version is 24.30 GB. The smallest option (Q2_K) is 5.16 GB.
- Which GPUs can run Mellum2.1 12B A2.5B Thinking?
52 consumer GPUs can run Mellum2.1 12B A2.5B Thinking at Q4_K_M (7.7 GB). Top options include AMD Radeon RX 6700 XT, AMD Radeon RX 6800, AMD Radeon RX 6800 XT, AMD Radeon RX 7600. 36 GPUs have plenty of headroom for comfortable inference.
- Which devices can run Mellum2.1 12B A2.5B Thinking?
50 devices with unified memory can run Mellum2.1 12B A2.5B Thinking at Q4_K_M (7.7 GB), including AMD Ryzen AI 9 HX 370 (Strix Point) Laptop, ASUS Ascent GX10, Asus ROG Flow Z13 (2025, Ryzen AI Max+ 395, 128 GB), Beelink GTR9 Pro (Ryzen AI Max+ 395, 128 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.