AliceAI Foundation 80B-A3B VRAM requirements
AliceAI Foundation 80B-A3B needs about 51.1 GB of VRAM at Q4_K_M with 8K context, so it needs a 64 GB setup such as 2× RTX 5090; a 24 GB card is too small. AliceAI Foundation 80B-A3B has 81.3B parameters, of which about 3B are used per token; all 512 experts still have to be in memory. With the same 8K context and one request it needs 84.0 GB at FP8 and 167 GB at FP16/BF16. The published weights take 151 GB (BF16). It does not fit on a 24 GB card at Q4_K_M; the smallest setup here that holds it with an 8K context is 2× RTX 5090 (64 GB).
51.1 GB at Q4_K_M, 8K context, one request
- FP8
- 84.0 GB
- FP16 / BF16
- 167 GB
- Published weights
- 151 GB
- Smallest setup, Q4_K_M
- 2× RTX 5090 (64 GB)
Worked example: AliceAI Foundation 80B-A3B with 32K tokens
Inputs: AliceAI Foundation 80B-A3B · Q4_K_M weights · 32,768 tokens of context · one request · FP16 KV cache
- Weights (81.3B at Q4_K_M)
- 45.8 GB
- KV cache (24 KB per token)
- 768 MB
- Buffers and runtime (0.5 GB + 10%)
- 5.16 GB
- Total
- 51.7 GB
The smallest setup here that holds it is 2× RTX 5090 (64 GB): split across the 2 cards, with the 0.5 GB runtime base on each of them, it takes 52.2 GB and leaves about 11.8 GB to spare. Change the inputs in the calculator
What makes AliceAI Foundation 80B-A3B's memory use different
36 of its 48 layers keep a fixed-size KDA recurrent state, so the cache grows by 24 KB per token where caching every layer would take 96 KB.
Per token of context it adds 24 KB of FP16 cache; Llama 3.1 70B (70.6B), the nearest-sized model here with plain full attention, adds 320 KB, so AliceAI Foundation 80B-A3B needs 8% as much.
As a mixture-of-experts model it reads about 1.69 GB of its 45.8 GB Q4_K_M weights per generated token (4%), so it writes like a much smaller model while needing memory for all of them.
How much VRAM does AliceAI Foundation 80B-A3B need?
AliceAI Foundation 80B-A3B needs 51.1 GB at Q4_K_M, 89.2 GB at Q8_0 and 167 GB at FP16 with 8,192 tokens of context and one request. Each total below is the weights plus the FP16 KV cache and the runtime overhead (0.5 GB plus 10%), and each row opens the calculator with that setting. What the GGUF names mean.
| Precision | Weights | Total | Smallest setup |
|---|---|---|---|
| As published (BF16) | 151 GB | 167 GB | B200 (180 GB) |
| FP16 / BF16 | 151 GB | 167 GB | B200 (180 GB) |
| FP8 / INT8 | 75.7 GB | 84.0 GB | Ryzen AI Max+ 395 (128 GB, 96 GB usable) |
| INT4 (AWQ / GPTQ) | 40.2 GB | 44.9 GB | M4 Pro Mac (64 GB, 48 GB usable) |
| GGUF Q8_0 | 80.4 GB | 89.2 GB | Ryzen AI Max+ 395 (128 GB, 96 GB usable) |
| GGUF Q6_K | 62.1 GB | 69.0 GB | A100 80GB |
| GGUF Q5_K_M | 53.7 GB | 59.7 GB | 2× RTX 5090 (64 GB) |
| GGUF Q4_K_M | 45.8 GB | 51.1 GB | 2× RTX 5090 (64 GB) |
| GGUF IQ4_XS | 41.2 GB | 46.0 GB | M4 Pro Mac (64 GB, 48 GB usable) |
| GGUF Q3_K_M | 37.0 GB | 41.4 GB | M4 Pro Mac (64 GB, 48 GB usable) |
| GGUF IQ3_XXS | 31.2 GB | 35.1 GB | M4 Pro Mac (64 GB, 48 GB usable) |
| GGUF Q2_K | 31.7 GB | 35.6 GB | M4 Pro Mac (64 GB, 48 GB usable) |
Fine-tuning AliceAI Foundation 80B-A3B? AliceAI Foundation 80B-A3B VRAM for LoRA, QLoRA and full training.
AliceAI Foundation 80B-A3B GGUF files on Hugging Face
The GGUF file found for AliceAI Foundation 80B-A3B, with the bytes the Hugging Face file list reported on 2026-09-29. Each total is that file plus the FP16 KV cache and 0.5 GB + 10%, one request. "vs estimate" compares the file with the bits-per-weight size the table above uses.
| Quant | File | Size | vs estimate | 8K | 32K | 128K |
|---|---|---|---|---|---|---|
| Q4_K_M | Yamada114514/ | 45.1 GB48,421,862,304 bytes | −2% | 50.3 GB | 50.9 GB | 53.4 GB |
GB here is GiB (1024³ bytes), as everywhere on this site; Hugging Face shows the Q4_K_M file as 48.4 GB, in 10⁹ bytes. AliceAI-Foundation-80B-A3B-Base-Q4_K_M.gguf is a community conversion: it needs the llama.cpp patch in the same repo (aliceai-llama.cpp.patch); stock llama.cpp, Ollama and LM Studio cannot load the aliceai architecture. The uploader reports a peak resident memory of 45.58 GiB on an M5 Max Mac with 128 GB, with -c 4096, one slot and an FP16 KV cache, on inputs of 21 to 336 tokens.
Which AliceAI Foundation 80B-A3B GGUF fits a 16, 24, 32, 48 or 80 GB GPU or a Mac
From the file sizes above: the largest file that fits each machine with an 8K context and 0.5 GB left free, the longest context it then has room for, and the longest context for Q4_K_M (45.1 GB). One request, FP16 KV cache. The Q4_K_M file is 721 MB under its bits-per-weight estimate; the estimate tables on this page count the estimate, so they show less context.
| Hardware | Largest GGUF, 8K | Its longest context | Q4_K_M, longest context |
|---|---|---|---|
| RTX 5060 Ti 16GB | None fits | — | No |
| RTX 4090 (24 GB) | None fits | — | No |
| RTX 5090 (32 GB) | None fits | — | No |
| L40S (48 GB) | None fits | — | No |
| H100 SXM (80 GB) | Q4_K_M, 45.1 GB | 256K (full) | 256K (full) |
| M4 Pro Mac (64 GB, 48 GB usable) | None fits | — | No |
| M4 Max Mac (128 GB, 96 GB usable) | Q4_K_M, 45.1 GB | 256K (full) | 256K (full) |
AliceAI Foundation 80B-A3B at 8K, 32K, 128K and 256K (full) tokens of context
12 of its 48 layers use full attention and 36 are KDA recurrent layers with no growing cache. Each extra token of context adds 24 KB of FP16 cache per request. The last column is the smallest setup here, one card or a group, that holds it at Q4_K_M. How the KV cache works.
| Context | KV cache, FP16 | Total at Q4_K_M | Total at Q8_0 | Smallest setup, Q4_K_M |
|---|---|---|---|---|
| 8K tokens | 192 MB | 51.1 GB | 89.2 GB | 2× RTX 5090 (64 GB) |
| 32K tokens | 768 MB | 51.7 GB | 89.8 GB | 2× RTX 5090 (64 GB) |
| 128K tokens | 3.00 GB | 54.2 GB | 92.3 GB | 2× RTX 5090 (64 GB) |
| 256K tokens | 6.00 GB | 57.5 GB | 95.6 GB | 2× RTX 5090 (64 GB) |
How fast AliceAI Foundation 80B-A3B writes
Tokens per second for one request with 8,192 tokens of context, estimated from memory bandwidth and the parameters read per token. A dash means it does not fit on one card. Try other settings in the speed calculator.
| Hardware | Bandwidth | Q4_K_M | FP8 |
|---|---|---|---|
| RTX 3060 12GB | 360 GB/s | — | — |
| RTX 4090 | 1,008 GB/s | — | — |
| RTX 5090 | 1,792 GB/s | — | — |
| M4 Max Mac (128 GB) | 546 GB/s | 72–127 | 48–82 |
| M3 Ultra Mac Studio (512 GB) | 819 GB/s | 103–184 | 69–120 |
| H100 SXM | 3,350 GB/s | 285–587 | — |
| H200 | 4,800 GB/s | 345–746 | 269–545 |
Longest context on one GPU
How many tokens of context AliceAI Foundation 80B-A3B fits on a single card with one request, an FP16 KV cache and 0.5 GB left free. "Full" means the model's whole context window fits.
| GPU | Memory | Q4_K_M | Q8_0 | FP8 |
|---|---|---|---|---|
| RTX 3060 | 12 GB | No | No | No |
| RTX 4060 Ti 16GB | 16 GB | No | No | No |
| RTX 3090 / 4090 | 24 GB | No | No | No |
| RTX 5090 | 32 GB | No | No | No |
| A100 40GB | 40 GB | No | No | No |
| Mac, 64 GB unified memory | 48 GB | No | No | No |
| L40S / RTX 6000 Ada | 48 GB | No | No | No |
| A100 / H100 80GB | 80 GB | 256K (full) | No | No |
| Mac, 128 GB unified memory | 96 GB | 256K (full) | 252K | 256K (full) |
| H200 | 141 GB | 256K (full) | 256K (full) | 256K (full) |
| B200 | 180 GB | 256K (full) | 256K (full) | 256K (full) |
Model details
- Parameters
- 81.3B (81,286,433,408)
- Experts
- 512 routed experts, all loaded
- Active per token
- 3B
- Layers
- 12 of its 48 layers use full attention and 36 are KDA recurrent layers with no growing cache
- Attention cache
- 2 KV heads × 256
- Context length
- 262,144 tokens
- Published weights
- 151 GB (BF16)
- On Hugging Face
- yandex/
AliceAI-Foundation-80B-A3B-Base - Reported benchmarks
- None published in the model card for GPQA, MMLU-Pro or LiveCodeBench.
Why this estimate looks this way
At Q4_K_M and an 8K-token context, AliceAI Foundation 80B-A3B uses 45.8 GB for weights, 192 MB for its FP16 KV cache and 5.10 GB for estimated runtime overhead, totaling 51.1 GB. The overhead is a 0.5 GB base plus 10% of weights and cache.
This is a mixture-of-experts model: all 512 experts contribute to the 81.3B parameters held in memory, even though only about 3B parameters run per token. Using only active parameters would understate VRAM.
Its attention layout matters for long context: 12 of its 48 layers use full attention and 36 are KDA recurrent layers with no growing cache. The FP16 cache grows by about 24 KB per additional token per request.
The 36 KDA layers keep a fixed recurrent state, which this estimate does not include; only the 12 gated-attention layers add a context-growing KV cache. Real serving memory may be higher.
The architecture and published checkpoint size come from the model's config.json and weight files. These are estimates rather than measured peak memory; inference engines can reserve extra buffers or preallocate the full configured cache.
AliceAI Foundation 80B-A3B next to similar-sized models
The three models here closest to it in parameter count, at Q4_K_M with 8,192 tokens of context.
| Model | Parameters | Total | KV per token | Smallest setup |
|---|---|---|---|---|
| AliceAI Foundation 80B-A3B | 81.3B, 3B active | 51.1 GB | 24 KB | 2× RTX 5090 (64 GB) |
| Qwen3-Coder-Next | 79.7B, 3B active | 50.1 GB | 24 KB | 2× RTX 5090 (64 GB) |
| Kolibri-1 | 78.1B, 3.5B active | 49.2 GB | 20 KB | 2× RTX 5090 (64 GB) |
| Llama 3.1 70B | 70.6B | 47.0 GB | 320 KB | M4 Pro Mac (64 GB, 48 GB usable) |
AliceAI Foundation 80B-A3B VRAM questions
How much VRAM do I need to run AliceAI Foundation 80B-A3B locally?
About 51.1 GB at Q4_K_M with an 8K context and one request, by the bits-per-weight estimate, and 167 GB at BF16. The Q4_K_M GGUF AliceAI-Foundation-80B-A3B-Base-Q4_K_M.gguf in Yamada114514/
Can I run AliceAI Foundation 80B-A3B on a 24 GB GPU?
No. The RTX 4090 (24 GB) alone cannot hold it: the smallest listed GGUF, Q4_K_M at 45.1 GB, needs 50.3 GB with an 8K context.
Which quantization of AliceAI Foundation 80B-A3B fits in 16 GB of VRAM?
The RTX 5060 Ti 16GB alone cannot hold it: the smallest listed GGUF, Q4_K_M at 45.1 GB, needs 50.3 GB with an 8K context.
Can I run AliceAI Foundation 80B-A3B on a Mac?
The M4 Pro Mac (64 GB, 48 GB usable) alone cannot hold it: the smallest listed GGUF, Q4_K_M at 45.1 GB, needs 50.3 GB with an 8K context. On the M4 Max Mac (128 GB, 96 GB usable), the largest listed GGUF that fits with an 8K context and 0.5 GB free is Q4_K_M (45.1 GB), with its full 256K-token context. By default macOS lets the GPU use about 75% of unified memory.
How much more VRAM does AliceAI Foundation 80B-A3B need for 32K or 128K tokens of context?
Its FP16 KV cache is 192 MB at 8K, 768 MB at 32K, 3.00 GB at 128K for one request, so 128K adds 2.81 GB over 8K, plus 10% overhead. A q8_0 cache
Why does AliceAI Foundation 80B-A3B use more VRAM than its GGUF file size?
The file holds only the weights. At Q4_K_M with an 8K context: 45.1 GB of weights (the file's 48,421,862,304 bytes) + 192 MB of FP16 KV cache + 5.03 GB of compute buffers and runtime (0.5 GB + 10%) = 50.3 GB.
Can I run AliceAI Foundation 80B-A3B in llama.cpp, Ollama or LM Studio?
Not in a stock build yet. The only GGUF listed here, AliceAI-Foundation-80B-A3B-Base-Q4_K_M.gguf in Yamada114514/
Does AliceAI Foundation 80B-A3B need less VRAM because only 3B parameters are active?
No. All 512 experts stay loaded, so the Q4_K_M file takes 45.1 GB. The active parameters decide speed: each generated token reads about 1.69 GB of them.
Badge for your model card
Paste it into a Hugging Face or GitHub README; it shows the Q4_K_M total above and links to this page. Q8_0 and FP16 badges.
Numbers read from the model files on Hugging Face on . Compare it with other models in the reproducible model and GPU report.