GigaChat3 10B A1.8B Base — Hardware Requirements & GPU Compatibility
Chat
GigaChat3 10B A1.8B Base is a 11.5B-parameter open language model from ai-sage. It supports a context window of up to 262,144 tokens. At Q4_K_M it needs about 7.52 GB of VRAM — see which GPUs and Macs can run it below.
Runs great on a 24 GB GPU
Runs from 8 GB (Q4_K_M · 16K)Best on 24 GB (Q4_K_M · ~270 tok/s · 256K)Macs from 16 GB (Q4_K_M · ~63 tok/s)
MIT
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
Run GigaChat3 10B A1.8B Base
One recommended way. Every other method, OS and quant is in the step-by-step guide.
Recommended
vllm serve ai-sage/GigaChat3-10B-A1.8B-base --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 3060 12GB (4 GB headroom) · Q4_K_M
- Generation speed
- ~128 tok/s
- generation speed
- Cost per 1M output tokens
- $0.15
- per 1M output tokens
How Much VRAM Does GigaChat3 10B A1.8B Base Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | On RTX 4090 (reference) | |
|---|---|---|---|---|
| 5.5 GB | 6.2 GB | 2-bit · K-quant — 2-bit quantization with K-quant improvements | Fits: 256K context · ~323 tok/s | |
| 6.2 GB | 6.9 GB | 3-bit medium — 3-bit medium quantization | Fits: 256K context · ~302 tok/s | |
| 7.5 GB | 8.2 GB | 4-bit medium — 4-bit medium quantization — most popular sweet spot | Fits: 256K context · ~270 tok/s | |
| 8.8 GB | 9.5 GB | 5-bit medium — 5-bit medium quantization — good quality/size tradeoff | Fits: 256K context · ~244 tok/s | |
| 10.1 GB | 10.8 GB | 6-bit, very good — 6-bit quantization, very good quality | Fits: 256K context · ~223 tok/s | |
| 12.1 GB | 12.8 GB | 8-bit, near-lossless — 8-bit quantization, near-lossless | Fits: 256K context · ~196 tok/s | |
| 23.6 GB | 24.2 GB | 16-bit · bfloat — Brain floating point 16 — preferred for training | Fits: 16K context · ~117 tok/s |
Selected Q4_K_M
4-bit medium quantization7.7GBat 16K context
- Weights
- 6.9 GB
- KV cache
- 0.5 GB
- Overhead
- 0.3 GB
Fits RTX 4090 (reference) · 16.3 GB free · ~270 tok/s
≥50 tok/s needs ≥124 GB/s · RTX 5090 ~369, RTX 3090 Ti ~270
Which Hardware Can Run GigaChat3 10B A1.8B Base?
Q4_K_M · 7.5 GBGigaChat3 10B A1.8B Base (Q4_K_M) requires 7.5 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 10+ GB is recommended. Using the full 262K context window can add up to 7.5 GB, bringing total usage to 15.0 GB. 52 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti, NVIDIA GeForce RTX 3080. 151 devices with unified memory can run GigaChat3 10B A1.8B Base, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB), NVIDIA Jetson Orin Nano 8GB (Super).
≤ 11 GB
≤ 11 GB tier →fits · 0.3 GB free · up to 16K ctx
16 GPUs · 1 device family
RTX 3080~225Simulate output on RTX 3080RTX 2080 Ti~194Simulate output on RTX 2080 TiRTX 3070 Ti~192Simulate output on RTX 3070 TiOrin Nano~42Simulate output on NVIDIA Jetson Orin Nano 8GB (Super)20 GB
fits · 12.3 GB free · up to 256K ctx
1 GPU
RX 7900 XT~120Simulate output on RX 7900 XT32–40 GB
fits · 24.3 GB free · up to 256K ctx
1 GPU · 3 Mac classes · 4 device families
RTX 5090~369Simulate output on RTX 5090M5 Max36 GB~102Simulate output on Mac Studio M5 Max (36 GB)M532 GB~52Simulate output on MacBook Air 13" M5 (32 GB)Orin~79Simulate output on NVIDIA Jetson AGX Orin 32GBSnapdragon X Elite (X1E-80-100)~51Simulate output on Snapdragon X Elite Copilot+ PC
By VRAM tier · typical GPUs and Macs at Q4_K_M
Similar models
6| Model | Score | VRAM | tok/s | Why it is similar |
|---|---|---|---|---|
| Gemma 3 12B ITGoogle | 74.2 | 10.9 GB | ~82 | same size |
| Gemma 4 12BGoogle | — | 9 GB | ~80 | same size |
| Qwen2.5 Omni 7BAlibaba | — | 9.9 GB | ~93 | same size |
| Llama 3.2 11B Vision InstructMeta | — | 9.8 GB | ~93 | same size |
| Mistral Nemo Base 2407Mistral AI | — | 11 GB | ~81 | same size |
| Falcon 11BTII UAE | — | 8.5 GB | ~89 | same size |
Specifications
- Publisher
- ai-sage
- Parameters
- 11.5B
- Active parameters
- 2.3B per token
- Architecture
- DeepseekV3ForCausalLM
- Context Length
- 262,144 tokens
- Vocabulary Size
- 128,256
- Release Date
- 2025-11-19
- License
- MIT
Frequently Asked Questions
How much VRAM does GigaChat3 10B A1.8B Base need?
GigaChat3 10B A1.8B Base requires 7.5 GB of VRAM at Q4_K_M, or 23.6 GB at BF16. Full 262K context adds up to 7.5 GB (15.0 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 11.5B × 4.8 bits ÷ 8 = 6.9 GB
KV Cache + Overhead ≈ 0.6 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 ≈ 8.1 GB (at full 262K context)
VRAM usage by quantization
What's the best quantization for GigaChat3 10B A1.8B Base?
For GigaChat3 10B A1.8B Base, Q4_K_M (7.5 GB) offers the best balance of quality and VRAM usage. Q5_K_M (8.8 GB) provides better quality if you have the VRAM. The smallest option is Q2_K at 5.5 GB.
VRAM requirement by quantization
★ Recommended — best balance of quality and VRAM usage.
Can I run GigaChat3 10B A1.8B Base on a Mac?
Yes — Mac Mini M2 Pro (16 GB) and 133 other Macs can run GigaChat3 10B A1.8B Base. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 7.5 GB of usable unified memory (RAM minus macOS overhead).
Can I run GigaChat3 10B A1.8B Base locally?
Yes — GigaChat3 10B A1.8B Base can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is GigaChat3 10B A1.8B Base?
At Q4_K_M, GigaChat3 10B A1.8B Base can reach ~182 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~270 tok/s. 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 → 1.5 GB active ÷ (8000 × 0.65) = 0.287 ms, plus 26 layers × 0.055 ms = 1.430 ms, so 1000 ÷ 1.717 ms = ~582 tok/s
Estimated speed at Q4_K_M (7.5 GB)
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
What's the download size of GigaChat3 10B A1.8B Base?
At Q4_K_M, the download is about 6.89 GB. The full-precision BF16 version is 22.96 GB. The smallest option (Q2_K) is 4.88 GB.
Which GPUs can run GigaChat3 10B A1.8B Base?
52 consumer GPUs can run GigaChat3 10B A1.8B Base at Q4_K_M. Top options include AMD Radeon RX 6700 XT, AMD Radeon RX 6800, AMD Radeon RX 6800 XT, AMD Radeon RX 7600. 36 GPUs have plenty of headroom for comfortable inference.
Which devices can run GigaChat3 10B A1.8B Base?
151 devices with unified memory can run GigaChat3 10B A1.8B Base at Q4_K_M (7.5 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.