Covenant 72B Chat — Hardware Requirements & GPU Compatibility
Chat
Covenant 72B Chat is a 72.7B-parameter open language model from 1Covenant. It supports a context window of up to 8,192 tokens. At Q4_K_M it needs about 44.62 GB of VRAM — see which GPUs and Macs can run it below.
Covenant 72B Chat runs on a 64 GB+ Mac or a workstation GPU
Best local pick → Mac Pro M2 Ultra (64 GB) · Q4_K_M · 46.6 GB · ~13 tok/s
Macs from 48 GB (Q3_K_M · ~12 tok/s)
Apache 2.0
Not set up for it?
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
Run Covenant 72B Chat
One recommended way. Every other method, OS and quant is in the step-by-step guide.Each way opens a step-by-step guide for your OS and quant.
Recommended
vllm serve 1Covenant/Covenant-72B-Chat --max-model-len 8192vLLM repo with its own quantisation
Download: Weights on Hugging Face
Ways to runOther ways
each opens step-by-stepRun in cloud
Runs on H100 (80 GB) (33 GB headroom) · Q4_K_M
- Generation speed
- ~49 tok/s
- generation speed
- Cost per 1M output tokens
- $15.31
- per 1M output tokens
How Much VRAM Does Covenant 72B Chat Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | On RTX 4090 (reference) | |
|---|---|---|---|---|
| 31.9 GB | — | 2-bit · K-quant — 2-bit quantization with K-quant improvements | Does not fit: needs 31.9 GB | |
| 36.4 GB | — | 3-bit medium — 3-bit medium quantization | Does not fit: needs 36.4 GB | |
| 44.6 GB | — | 4-bit medium — 4-bit medium quantization — most popular sweet spot | Does not fit: needs 44.6 GB | |
| 52.8 GB | — | 5-bit medium — 5-bit medium quantization — good quality/size tradeoff | Does not fit: needs 52.8 GB | |
| 61.0 GB | — | 6-bit, very good — 6-bit quantization, very good quality | Does not fit: needs 61 GB | |
| 73.7 GB | — | 8-bit, near-lossless — 8-bit quantization, near-lossless | Does not fit: needs 73.7 GB | |
| 146.5 GB | — | 16-bit · bfloat — Brain floating point 16 — preferred for training | Does not fit: needs 146.5 GB |
Selected Q4_K_M
4-bit medium quantization46.6GBat 8K context
- Weights
- 43.7 GB
- KV cache
- 2.7 GB
- Overhead
- 0.3 GB
Doesn't fit RTX 4090 (reference) · needs 22.6 GB more
Or .
Which Hardware Can Run Covenant 72B Chat?
Q4_K_M · 44.6 GBCovenant 72B Chat (Q4_K_M) requires 44.6 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 59+ GB is recommended. Using the full 8K context window can add up to 2.0 GB, bringing total usage to 46.6 GB. No consumer GPU has enough memory. 52 devices with unified memory can run Covenant 72B Chat, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB), Mac Pro M2 Ultra (64 GB).
64–80 GB
fits · 17.4 GB free · up to 8K ctx
3 Mac classes · 1 device family
M2 Ultra*64 GB~13Simulate output on Mac Pro M2 Ultra (64 GB)M5 Max64 GB~10Simulate output on Mac Studio M5 Max (64 GB)Orin~3Simulate output on NVIDIA Jetson AGX Orin 64GB
By VRAM tier · typical GPUs and Macs at Q4_K_M
Similar models
6| Model | Score | VRAM | tok/s | Why it is similar |
|---|---|---|---|---|
| Llama 3.3 70B InstructMeta | 80.0 | 51.3 GB | —needs > 24 GB | same size |
| NVIDIA Nemotron 3.5 Bolt 74B A9B BF16NVIDIA | — | 54.2 GB | —needs > 24 GB | same size |
| Qwen2.5 VL 72B InstructAlibaba | — | 49.7 GB | —needs > 24 GB | same size |
| Qwen2 VL 72B InstructAlibaba | — | 49.7 GB | —needs > 24 GB | same size |
| Qwen1.5 72B ChatAlibaba | — | 86.6 GB | —needs > 24 GB | same size |
| CodeLlama 70B Instruct HFMeta | — | 23.2 GB | ~29 | same size |
Specifications
- Publisher
- 1Covenant
- Parameters
- 72.7B
- Architecture
- LlamaForCausalLM
- Context Length
- 8,192 tokens
- Vocabulary Size
- 262,144
- Release Date
- 2026-02-09
- License
- Apache 2.0
Frequently Asked Questions
How much VRAM does Covenant 72B Chat need?
Covenant 72B Chat requires 44.6 GB of VRAM at Q4_K_M, or 146.5 GB at BF16. Full 8K context adds up to 2.0 GB (46.6 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 72.7B × 4.8 bits ÷ 8 = 43.6 GB
KV Cache + Overhead ≈ 1 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 ≈ 3 GB (at full 8K context)
VRAM usage by quantization
Can NVIDIA GeForce RTX 5090 run Covenant 72B Chat?
Yes, at Q2_K (31.9 GB) or lower. Higher quantizations like Q3_K_M (36.4 GB) exceed the NVIDIA GeForce RTX 5090's 32 GB.
What's the best quantization for Covenant 72B Chat?
For Covenant 72B Chat, Q4_K_M (44.6 GB) offers the best balance of quality and VRAM usage. Q5_K_M (52.8 GB) provides better quality if you have the VRAM. The smallest option is Q2_K at 31.9 GB.
VRAM requirement by quantization
★ Recommended — best balance of quality and VRAM usage.
Can I run Covenant 72B Chat on a Mac?
Yes — Mac Studio M5 Max (48 GB) and 55 other Macs can run Covenant 72B Chat, the smallest at Q3_K_M. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 44.6 GB of usable unified memory (RAM minus macOS overhead), from Mac Pro M2 Ultra (64 GB).
Can I run Covenant 72B Chat locally?
Yes — Covenant 72B Chat can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is Covenant 72B Chat?
At Q4_K_M, Covenant 72B Chat can reach ~108 tok/s on AMD Instinct MI350X. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = (bandwidth GB/s ÷ model GB) × efficiency
Example: NVIDIA B200 → 8000 ÷ 44.6 × 0.65 = ~117 tok/s
Estimated speed at Q4_K_M (44.6 GB)
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
What's the download size of Covenant 72B Chat?
At Q4_K_M, the download is about 43.65 GB. The full-precision BF16 version is 145.49 GB. The smallest option (Q2_K) is 30.92 GB.
Which GPUs can run Covenant 72B Chat?
No single consumer GPU has enough VRAM to run Covenant 72B Chat at Q4_K_M (44.6 GB). Multi-GPU or professional hardware is required.
Which devices can run Covenant 72B Chat?
52 devices with unified memory can run Covenant 72B Chat at Q4_K_M (44.6 GB), including ASUS Ascent GX10, Beelink GTR9 Pro (Ryzen AI Max+ 395, 128 GB), Framework Desktop (Ryzen AI Max+ 395, 128 GB), GMKtec EVO-X2 (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.