Dhara 70M — Hardware Requirements & GPU Compatibility
ChatCode
Dhara 70M is a 71M-parameter open language model from codelion. It supports a context window of up to 1,024 tokens. At BF16 it needs about 0.49 GB of VRAM — see which GPUs and Macs can run it below.
Runs great on a 24 GB GPU
Runs from 8 GB (BF16)Best on 24 GB (BF16 · ~1337 tok/s)Macs from 8 GB (BF16 · ~146 tok/s)
Apache 2.0
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
Run Dhara 70M
One recommended way. Every other method, OS and quant is in the step-by-step guide.
Recommended
vllm serve codelion/dhara-70m --max-model-len 1024vLLM repo with its own quantisation
Download: Weights on Hugging Face
Other ways
each opens step-by-stepRun in cloud
Fits on RTX 3060 12GB (11 GB headroom) · BF16
- Generation speed
- ~478 tok/s
- generation speed
- Cost per 1M output tokens
- $0.048
- per 1M output tokens
How Much VRAM Does Dhara 70M Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | On RTX 4090 (reference) | |
|---|---|---|---|---|
| 0.5 GB | — | 16-bit · bfloat — Brain floating point 16 — preferred for training | Fits: — context · ~1337 tok/s |
Selected BF16
Brain floating point 160.5GBat 1K context
- Weights
- 0.1 GB
- KV cache
- 0.0 GB
- Overhead
- 0.3 GB
Fits RTX 4090 (reference) · 23.5 GB free · ~1337 tok/s
≥50 tok/s needs ≥38 GB/s · RTX 5090 ~2377, RTX 3090 Ti ~1337
Which Hardware Can Run Dhara 70M?
BF16 · 0.5 GBDhara 70M (BF16) requires 0.5 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 1+ GB is recommended. 52 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti. 163 devices with unified memory can run Dhara 70M, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB).
≤ 11 GB
≤ 11 GB tier →fits · 7.5 GB free · up to — ctx
16 GPUs · 1 Mac class · 2 device families
RTX 3080~1009Simulate output on RTX 3080RTX 2080 Ti~817Simulate output on RTX 2080 TiRTX 3070 Ti~807Simulate output on RTX 3070 TiM3*8 GB~146Simulate output on iMac 24" M3 (8 GB)Orin Nano~135Simulate output on NVIDIA Jetson Orin Nano 8GB (Super)iPhone~97Simulate output on iPhone 1720 GB
fits · 19.5 GB free · up to — ctx
1 GPU
RX 7900 XT~980Simulate output on RX 7900 XT32–40 GB
fits · 31.5 GB free · up to — ctx
1 GPU · 3 Mac classes · 4 device families
RTX 5090~2377Simulate output on RTX 5090M5 Max36 GB~657Simulate output on Mac Studio M5 Max (36 GB)M532 GB~219Simulate output on MacBook Air 13" M5 (32 GB)Orin~272Simulate output on NVIDIA Jetson AGX Orin 32GBSnapdragon X Elite (X1E-80-100)~179Simulate output on Snapdragon X Elite Copilot+ PC
By VRAM tier · typical GPUs and Macs at BF16
Similar models
1| Model | Score | VRAM | tok/s | Why it is similar |
|---|---|---|---|---|
| Llama 68MJackFram | — | 0 GB | ~16380 | same size |
Specifications
- Publisher
- codelion
- Parameters
- 71M
- Architecture
- DharaForMaskedDiffusion
- Context Length
- 1,024 tokens
- Vocabulary Size
- 50,304
- Release Date
- 2025-12-24
- License
- Apache 2.0
Frequently Asked Questions
How much VRAM does Dhara 70M need?
Dhara 70M requires 0.5 GB of VRAM at BF16.
VRAM = Weights + KV Cache + Overhead
Weights = 71M × 16 bits ÷ 8 = 0.1 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.
VRAM usage by quantization
Can I run Dhara 70M on a Mac?
Yes — MacBook Air 13" M3 (8 GB) and 145 other Macs can run Dhara 70M. Apple Silicon uses unified memory, so the model shares RAM with the system. At BF16 you need at least 0.5 GB of usable unified memory (RAM minus macOS overhead).
Can I run Dhara 70M locally?
Yes — Dhara 70M can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is Dhara 70M?
At BF16, Dhara 70M can reach ~9796 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~1337 tok/s. 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 ÷ 0.49 × 0.65 = ~10612 tok/s
Estimated speed at BF16 (0.5 GB)
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
What's the download size of Dhara 70M?
At BF16, the download is about 0.14 GB.
Which GPUs can run Dhara 70M?
52 consumer GPUs can run Dhara 70M at BF16. Top options include AMD Radeon RX 6700 XT, AMD Radeon RX 6800, AMD Radeon RX 6800 XT. 52 GPUs have plenty of headroom for comfortable inference.
Which devices can run Dhara 70M?
163 devices with unified memory can run Dhara 70M at BF16 (0.5 GB), including AMD Ryzen AI 9 HX 370 (Strix Point) Laptop, ASUS Ascent GX10, Beelink GTR9 Pro (Ryzen AI Max+ 395, 128 GB), Framework Desktop (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.