internlm2-math-plus-mixtral8x22b needs about 66.9 GB of memory at IQ3_M (a 64.5 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 25 quants, from 29.6 GB (IQ1_S) to 281.3 GB. Longer context needs more: at 32K tokens the KV cache alone adds about 7.5 GB.
InternLM2-Math-Plus Mixtral8x22B – large‑scale math reasoning LLM
InternLM2-Math-Plus Mixtral8x22B is a 141‑billion‑parameter model built on the Mistral family. It supports 65536‑token context and is optimized for mathematical reasoning tasks. The model is available in multiple GGUF quantizations.
Good at
- Advanced math reasoning
- Code‑intpreter tasks
- Formal theorem proving
Limits
- No vision support
- Requires large VRAM for full precision
- License is other
Best for
- Math problem solving
- Educational tutoring
- Research prototyping
Who it’s for. Users needing high‑accuracy math reasoning in a large‑context setting. Suitable for developers and researchers who can allocate sufficient GPU memory.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| IQ1_S | 29.6 GB | 32.0 GB | Too big | Get file → | |
| IQ1_M | 32.7 GB | 35.1 GB | Too big | Get file → | |
| IQ2_XXS | 37.9 GB | 40.3 GB | Too big | Get file → | |
| IQ2_XS | 42.0 GB | 44.4 GB | Too big | Get file → | |
| IQ2_S | 42.6 GB | 45.0 GB | Too big | Get file → | |
| IQ2_M | 46.7 GB (2 parts) | 49.1 GB | Too big | Get file → | |
| Q2_K_S | 48.1 GB (2 parts) | 50.5 GB | Too big | Get file → | |
| Q2_K | 52.1 GB (3 parts) | 54.5 GB | Too big | Get file → | |
| IQ3_XXS | 54.9 GB (3 parts) | 57.3 GB | Too big | Get file → | |
| IQ3_XS | 58.2 GB (3 parts) | 60.6 GB | Too big | Get file → | |
| IQ3_S | 61.5 GB (3 parts) | 63.9 GB | Too big | Get file → | |
| Q3_K_S | 61.5 GB (3 parts) | 63.9 GB | Too big | Get file → | |
| IQ3_M | 64.5 GB (3 parts) | 66.9 GB | Too big | Get file → | |
| Q3_K | 67.8 GB (3 parts) | 70.2 GB | Too big | Get file → | |
| Q3_K_L | 72.6 GB (3 parts) | 75.0 GB | Too big | Get file → | |
| IQ4_XS | 75.5 GB (3 parts) | 77.9 GB | Too big | Get file → | |
| IQ4_NL | 79.8 GB (4 parts) | 82.2 GB | Too big | Get file → | |
| Q4_K_S | 80.5 GB (4 parts) | 82.9 GB | Too big | Get file → | |
| Q4_K | 85.6 GB (4 parts) | 88.0 GB | Too big | Get file → | |
| Q5_K_S | 97.0 GB (4 parts) | 99.4 GB | Too big | Get file → | |
| Q5_K | 100.0 GB (4 parts) | 102.4 GB | Too big | Get file → | |
| Q6_K | 115.5 GB (4 parts) | 117.9 GB | Too big | Get file → | |
| Q8_0 | 149.4 GB (5 parts) | 151.8 GB | Too big | Get file → | |
| BF16 | 281.3 GB (9 parts) | 283.6 GB | Too big | Get file → | |
| — | 281.3 GB (12 parts) | 283.6 GB | Too big | Get file → |
For 8 GB VRAM, use IQ1_S (29.65 GB file, ~32 GB memory). For 12 GB VRAM, IQ1_M (32.74 GB file, ~35.1 GB memory). For 16 GB VRAM, IQ2_S (42.6 GB file, ~45 GB memory).
What GPU do you need for internlm2-math-plus-mixtral8x22b?
These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.
- 4 GB VRAM: no quant fits at an 8K context.
- 6 GB VRAM: no quant fits at an 8K context.
- 8 GB VRAM: no quant fits at an 8K context.
- 12 GB VRAM: no quant fits at an 8K context.
- 16 GB VRAM: no quant fits at an 8K context.
- 24 GB VRAM: no quant fits at an 8K context.
- 32 GB VRAM: the best quant that fits is IQ1_S (29.6 GB file, about 32.0 GB in use).
- 48 GB VRAM: the best quant that fits is IQ2_S (42.6 GB file, about 45.0 GB in use).
Which internlm2-math-plus-mixtral8x22b quant should you choose?
IQ3_M (64.5 GB) is the middle option in this repository. A reasonable choice when a 4-bit file will not fit, mostly on larger models. Quality loss is noticeable but often acceptable.
If memory is tight, the smallest file is IQ1_S at 29.6 GB, with the quality trade-offs that come with fewer bits. Use it only when memory is very tight. Expect a clear drop in quality, which is usually worse on small models and easier to live with on large ones.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Frequently asked questions about internlm2-math-plus-mixtral8x22b
How much VRAM does internlm2-math-plus-mixtral8x22b need?
About 66.9 GB at IQ3_M with an 8K context window. That is the 64.5 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.
Can I run internlm2-math-plus-mixtral8x22b on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 32.0 GB, so you would need a larger GPU or CPU offloading.
What is the file size of internlm2-math-plus-mixtral8x22b GGUF?
It depends on the quant: IQ1_S is 29.6 GB, IQ1_M is 32.7 GB, IQ2_XXS is 37.9 GB, IQ2_XS is 42.0 GB, IQ2_S is 42.6 GB, IQ2_M is 46.7 GB and 19 more.
What context length does internlm2-math-plus-mixtral8x22b support?
The GGUF header lists a trained context length of 65,536 tokens (64K). The KV cache grows with the context you use: about 1.9 GB at 8K and 7.5 GB at 32K.
Where can I download internlm2-math-plus-mixtral8x22b GGUF?
The files are on Hugging Face in the legraphista/internlm2-math-plus-mixtral8x22b-IMat-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.
What license does internlm2-math-plus-mixtral8x22b use?
Hugging Face lists the license as other. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.
Can I run InternLM2-Math-Plus Mixtral8x22B on 8 GB VRAM?
Use the IQ1_S quantization; it requires about 32 GB of VRAM at 8k context, which is beyond 8 GB, so it is not feasible.
Does the model support image input?
No, the model does not support vision capabilities.
What is the context length of this model?
The model was trained with a context length of 65536 tokens.