INTELLECT 3 REAP 50 Heretic — Hardware Requirements & GPU Compatibility
Chat
INTELLECT 3 REAP 50 Heretic is a 57.0B-parameter open language model from jtl11. It supports a context window of up to 131,072 tokens. At Q4_K_M it needs about 34.64 GB of VRAM — see which GPUs and Macs can run it below.
Doesn't fit a 24 GB GPU
Smallest consumer fit → Snapdragon X Elite Copilot+ PC · Q2_K · 25.6 GB · ~15 tok/s
Macs from 32 GB (Q2_K · ~35 tok/s)
Apache 2.0
Not set up for it?
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
Run INTELLECT 3 REAP 50 Heretic
One recommended way. Every other method, OS and quant is in the step-by-step guide.
Recommended
vllm serve jtl11/INTELLECT-3-REAP-50-heretic --max-model-len 16384vLLM repo with its own quantisation
Download: Weights on Hugging Face
Other ways
each opens step-by-stepRun in cloud
Fits on RTX A6000 (48 GB) (12 GB headroom) · Q4_K_M
- Generation speed
- ~57 tok/s
- generation speed
- Cost per 1M output tokens
- $1.62
- per 1M output tokens
How Much VRAM Does INTELLECT 3 REAP 50 Heretic Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | On RTX 4090 (reference) | |
|---|---|---|---|---|
| 24.7 GB | 26.6 GB | 2-bit · K-quant — 2-bit quantization with K-quant improvements | Does not fit: needs 25.6 GB | |
| 28.2 GB | 30.2 GB | 3-bit medium — 3-bit medium quantization | Does not fit: needs 29.1 GB | |
| 34.6 GB | 36.6 GB | 4-bit medium — 4-bit medium quantization — most popular sweet spot | Does not fit: needs 35.5 GB | |
| 41.0 GB | 43 GB | 5-bit medium — 5-bit medium quantization — good quality/size tradeoff | Does not fit: needs 42 GB | |
| 47.5 GB | 49.4 GB | 6-bit, very good — 6-bit quantization, very good quality | Does not fit: needs 48.4 GB | |
| 57.4 GB | 59.4 GB | 8-bit, near-lossless — 8-bit quantization, near-lossless | Does not fit: needs 58.3 GB | |
| 114.5 GB | 116.4 GB | 16-bit · bfloat — Brain floating point 16 — preferred for training | Does not fit: needs 115.4 GB |
Selected Q4_K_M
4-bit medium quantization35.5GBat 16K context
- Weights
- 34.2 GB
- KV cache
- 1.0 GB
- Overhead
- 0.3 GB
Doesn't fit RTX 4090 (reference) · needs 11.5 GB more
Which Hardware Can Run INTELLECT 3 REAP 50 Heretic?
Q4_K_M · 34.6 GBINTELLECT 3 REAP 50 Heretic (Q4_K_M) requires 34.6 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 46+ GB is recommended. Using the full 131K context window can add up to 8.1 GB, bringing total usage to 42.7 GB. No consumer GPU has enough memory. 67 devices with unified memory can run INTELLECT 3 REAP 50 Heretic, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB), Mac Studio M5 Max (48 GB).
32–40 GB
Q4_K_M · best fit Q3_K_M (29.1 GB)
1 GPU · 3 Mac classes · 1 device family
RTX 5090~128Simulate output on RTX 5090M5 Max36 GB~35Simulate output on Mac Studio M5 Max (36 GB)M532 GB~15Simulate output on MacBook Air 13" M5 (32 GB)Orin~21Simulate output on NVIDIA Jetson AGX Orin 32GB48 GB
48 GB tier →fits · 12.5 GB free · up to 128K ctx
2 Mac classes · 1 device family
M5 Max48 GB~37Simulate output on Mac Studio M5 Max (48 GB)M5 Pro48 GB~23Simulate output on Mac Mini M5 Pro (48 GB)Snapdragon X2 Elite Extreme (X2E-96-100)~18Simulate output on Snapdragon X2 Elite Extreme Copilot+ PC
By VRAM tier · typical GPUs and Macs at Q4_K_M
Similar models
2| Model | Score | VRAM | tok/s | Why it is similar |
|---|---|---|---|---|
| Qwen2 57B A14B InstructAlibaba | — | 35.7 GB | —needs > 24 GB | same size |
| Mixtral 34Bx2 MoE 60Bcloudyu | — | 40.8 GB | —needs > 24 GB | same size |
Specifications
- Publisher
- jtl11
- Parameters
- 57.0B
- Active parameters
- 12.4B per token
- Architecture
- Glm4MoeForCausalLM
- Context Length
- 131,072 tokens
- Vocabulary Size
- 151,552
- Release Date
- 2026-01-18
- License
- Apache 2.0
Frequently Asked Questions
How much VRAM does INTELLECT 3 REAP 50 Heretic need?
INTELLECT 3 REAP 50 Heretic requires 34.6 GB of VRAM at Q4_K_M, or 114.5 GB at BF16. Full 131K context adds up to 8.1 GB (42.7 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 57.0B × 4.8 bits ÷ 8 = 34.2 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 ≈ 8.5 GB (at full 131K context)
VRAM usage by quantization
Can NVIDIA GeForce RTX 5090 run INTELLECT 3 REAP 50 Heretic?
Yes, at Q3_K_M (28.2 GB) or lower. Higher quantizations like Q4_K_M (34.6 GB) exceed the NVIDIA GeForce RTX 5090's 32 GB.
What's the best quantization for INTELLECT 3 REAP 50 Heretic?
For INTELLECT 3 REAP 50 Heretic, Q4_K_M (34.6 GB) offers the best balance of quality and VRAM usage. Q5_K_M (41.0 GB) provides better quality if you have the VRAM. The smallest option is Q2_K at 24.7 GB.
VRAM requirement by quantization
★ Recommended — best balance of quality and VRAM usage.
Can I run INTELLECT 3 REAP 50 Heretic on a Mac?
Yes — Mac Studio M1 Max (32 GB) and 84 other Macs can run INTELLECT 3 REAP 50 Heretic, the smallest at Q2_K. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 34.6 GB of usable unified memory (RAM minus macOS overhead), from Mac Studio M5 Max (48 GB).
Can I run INTELLECT 3 REAP 50 Heretic locally?
Yes — INTELLECT 3 REAP 50 Heretic can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is INTELLECT 3 REAP 50 Heretic?
At Q4_K_M, INTELLECT 3 REAP 50 Heretic can reach ~93 tok/s on AMD Instinct MI350X. 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: just the active experts get read for each token, along with a fixed routing cost per layer.
Example: NVIDIA B200 → 7.6 GB active ÷ (8000 × 0.65) = 1.45 ms, plus 46 layers × 0.055 ms = 2.53 ms, so 1000 ÷ 3.98 ms = ~251 tok/s
Estimated speed at Q4_K_M (34.6 GB)
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
What's the download size of INTELLECT 3 REAP 50 Heretic?
At Q4_K_M, the download is about 34.21 GB. The full-precision BF16 version is 114.02 GB. The smallest option (Q2_K) is 24.23 GB.
Which GPUs can run INTELLECT 3 REAP 50 Heretic?
No single consumer GPU has enough VRAM to run INTELLECT 3 REAP 50 Heretic at Q4_K_M (34.6 GB). Multi-GPU or professional hardware is required.
Which devices can run INTELLECT 3 REAP 50 Heretic?
67 devices with unified memory can run INTELLECT 3 REAP 50 Heretic at Q4_K_M (34.6 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.