Deep Cogito32B~20 GB VRAM at Q4_K_M

Cogito v1 32B — VRAM, speed & local setup

Written by Jakub Rusinowski · Last updated

Flagship consumer Cogito variant. Scores 52.7% on AIME 2025 — higher than DeepSeek-R1 32B. Runs on a single RTX 3090 or 4090 (24 GB) at ~18 tokens/second. The best open reasoning model available on a single consumer GPU.

Cogito v1 32B needs about 20 GB of VRAM at Q4_K_M — quantized weights plus framework overhead, before any KV cache. On Apple Silicon that figure comes out of unified memory.

Logic88
Creative80
Coding84

VRAM and speed by quantization

Quoted against NVIDIA RTX 4090 (24 GB). Includes the KV cache at 8K context, so it reads higher than the headline figure.

QuantVRAMSpeed (est.)Fit
Q2_K
2.63 bpw
13.5 GB~57 tok/sFits
Q3_K_M
3.41 bpw
16.6 GB~46 tok/sFits
Q4_K_M
4.83 bpw
22.3 GB~34 tok/sTight
Q5_K_M
5.67 bpw
25.6 GB~4 tok/sOffload
Q6_K
6.56 bpw
29.2 GB~4 tok/sOffload
Q8_0
8.50 bpw
36.9 GB~3 tok/sOffload
F16
16.00 bpw
66.9 GB—Too big

Black marker = usable memory on the NVIDIA RTX 4090 (24 GB). Estimates from the memory-bandwidth roofline on the methodology page. Cogito v1 32B VRAM calculator →

Get Cogito v1 32B running

The cheapest catalogued GPU that runs Cogito v1 32B is the AMD Radeon RX 7900 XTX (24 GB).

Buy this hardware AMD Radeon RX 7900 XTX 24GB — 24 GB VRAM · 355 W board powerDeploy in the cloud now RTX 4090 on RunPod — from $0.34/hr · rate checked 2026-07

or compare on Vast.ai from $0.35/hr (typical low · varies)

As an Amazon Associate we earn from qualifying purchases. Cloud GPU links are referral links — we may earn a commission at no extra cost to you.

Affiliate disclosure: Some links on this page are affiliate links — if you buy through them, LLM Configurator may earn a commission at no extra cost to you. As an Amazon Associate, LLM Configurator earns from qualifying purchases.
AMD Radeon RX 7900 XTX 24GB
24 GB VRAM · 355 W board power
2026 prices are volatile — check the current listing.

How to run Cogito v1 32B

Install Ollama, then run:

ollama run cogito:32b
Weights on Hugging Face: deepcogito/Cogito-v1-32B-Instruct ↗

Specifications

Curated — A hand-written entry from before this catalogue recorded its sources. The figures are long-standing but their provenance is not on file.

Parameters
32 Billion
Context window
64,000
Architecture
Hybrid Reasoning Transformer
Provider
Deep Cogito
Licence
Apache 2.0
Practical deployment
Single consumer GPU
Specified at
Q4_K_M
System RAM
48 GB
Record updated
2026-03-20
LicenceApache-2.0Commercial use permitted

Commercial use permitted. No usage restrictions beyond attribution.

Quality and use cases

Scores as published by the model’s authors or an independent evaluator — quality, not throughput, and not measured by us.

Best forreasoningmathcodingrtx 4090
BenchmarkScoreProvenance
MMLU85.6 / 100 %reported
MATH78.3 / 100 %reported
AIME 202552.7 / 100 %reported
HumanEval83.1 / 100 %reported

Can I run Cogito v1 32B on my GPU?

Other Cogito v1 sizes

Cogito v1 32B — frequently asked questions

How much VRAM does Cogito v1 32B need?

About 20 GB at Q4_K_M — quantized weights plus framework overhead, before any KV cache. The cache grows with context length and is added on top; the table above folds it in. Apple Silicon counts unified memory toward the same figure.

Does Cogito v1 32B run on an RTX 4090 (24 GB)?

Yes. Cogito v1 32B needs about 20 GB at Q4_K_M, inside a 24 GB card, at an estimated 34 tokens/sec.

How do I run Cogito v1 32B locally?

Install Ollama and run `ollama run cogito:32b`. That pulls the weights and starts a local OpenAI-compatible endpoint; after the download nothing leaves the machine.

What other sizes does Cogito v1 come in?

Cogito v1 3B (3 GB), Cogito v1 8B (6 GB), Cogito v1 14B (9 GB), Cogito v1 32B (20 GB), Cogito v1 70B (43 GB). Every size shares the family's training and licence; the larger ones score higher and need proportionally more memory.