Kimi K2.6 Eagle3 — Hardware Requirements & GPU Compatibility
Chat
Kimi K2.6 Eagle3 is a 667M-parameter open language model from NVIDIA in the Kimi K2 family. It supports a context window of up to 262,144 tokens. At BF16 it needs about 1.69 GB of VRAM — see which GPUs and Macs can run it below.
207 downloads
3 likesSpecifications
- Publisher
- NVIDIA
- Family
- Kimi K2
- Parameters
- 667M
- Architecture
- Eagle3DeepseekV2ForCausalLM
- Context Length
- 262,144 tokens
- Vocabulary Size
- 163,840
- Release Date
- 2026-06-03
- License
- Other
Get Started
HuggingFace
How Much VRAM Does Kimi K2.6 Eagle3 Need?
Select a quantization to see compatible GPUs below.
| Quantization | Bits | VRAM | + Context | File Size | Quality |
|---|---|---|---|---|---|
| BF16 | 16.00 | 1.7 GB | 9.2 GB | 1.33 GB | Brain floating point 16 — preferred for training |
Which GPUs Can Run Kimi K2.6 Eagle3?
BF16 · 1.7 GBKimi K2.6 Eagle3 (BF16) requires 1.7 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 3+ GB is recommended. Using the full 262K context window can add up to 7.5 GB, bringing total usage to 9.2 GB. 35 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti.
Runs great
— Plenty of headroomWhich Devices Can Run Kimi K2.6 Eagle3?
BF16 · 1.7 GB33 devices with unified memory can run Kimi K2.6 Eagle3, including NVIDIA DGX H100, NVIDIA DGX A100 640GB.
Runs great
— Plenty of headroomRelated Models
Frequently Asked Questions
- How much VRAM does Kimi K2.6 Eagle3 need?
Kimi K2.6 Eagle3 requires 1.7 GB of VRAM at BF16. Full 262K context adds up to 7.5 GB (9.2 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 667M × 16 bits ÷ 8 = 1.3 GB
KV Cache + Overhead ≈ 0.4 GB (at 2K context + ~0.3 GB framework)
KV Cache + Overhead ≈ 7.9 GB (at full 262K context)
VRAM usage by quantization
BF161.7 GBBF16 + full context9.2 GB- Can I run Kimi K2.6 Eagle3 on a Mac?
Kimi K2.6 Eagle3 requires at least 1.7 GB at BF16, which exceeds the unified memory of most consumer Macs. You would need a Mac Studio or Mac Pro with a high-memory configuration.
- Can I run Kimi K2.6 Eagle3 locally?
Yes — Kimi K2.6 Eagle3 can run locally on consumer hardware. At BF16 quantization it needs 1.7 GB of VRAM. Popular tools include Ollama, LM Studio, and llama.cpp.
- How fast is Kimi K2.6 Eagle3?
At BF16, Kimi K2.6 Eagle3 can reach ~1725 tok/s on AMD Instinct MI300X. On NVIDIA GeForce RTX 4090: ~388 tok/s. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = (bandwidth GB/s ÷ model GB) × efficiency
Example: AMD Instinct MI300X → 5300 ÷ 1.7 × 0.55 = ~1725 tok/s
Estimated speed at BF16 (1.7 GB)
~1725 tok/s~388 tok/s~1289 tok/s~1066 tok/sReal-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
- What's the download size of Kimi K2.6 Eagle3?
At BF16, the download is about 1.33 GB.
- Which GPUs can run Kimi K2.6 Eagle3?
35 consumer GPUs can run Kimi K2.6 Eagle3 at BF16 (1.7 GB). Top options include AMD Radeon RX 6700 XT, AMD Radeon RX 6800, AMD Radeon RX 6800 XT. 35 GPUs have plenty of headroom for comfortable inference.
- Which devices can run Kimi K2.6 Eagle3?
33 devices with unified memory can run Kimi K2.6 Eagle3 at BF16 (1.7 GB), including Mac Mini M4 (16 GB), Mac Mini M4 (32 GB), Mac Mini M4 Pro (24 GB), Mac Mini M4 Pro (48 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.