Mistral Large 4 1T A52B — Hardware Requirements & GPU Compatibility
VisionFunctions
Mistral Large 4 is Mistral AI's largest model to date, a natively multimodal Mixture-of-Experts model with about 1 trillion total parameters and roughly 52 billion active per token (49 billion excluding embeddings and output layers). It uses 128 routed experts with 4 active per token plus one shared expert, and a Pixtral-style vision encoder for image input. A single set of weights covers instruction following, reasoning with configurable effort, coding, and agentic work with native function calling, and it is fluent in more than 160 languages. All experts must still fit in memory even though only a fraction is active. Context length is 1M tokens. It is released under the Mistral Ecosystem Open-Weight License, a modified MIT license for commercial and non-commercial use, though running a Model-as-a-Service business requires a separate commercial license. It was published in September 2026. This is a frontier-scale model that needs a multi-GPU server or a 512 GB unified-memory machine even when quantized; an NVFP4 checkpoint targeting Blackwell GPUs is also available.
Mistral Large 4 1T A52B runs on a multi-GPU server
Q4_K_M needs ~629.3 GB · about 5× H200 SXM 141 GB · or Mac Studio (M3 Ultra, 512GB) at Q2_K (~19 tok/s)
Macs from 512 GB (Q2_K · ~19 tok/s)
Score 118#18 of 104 open#16 of 101 that fit ≤640 GBOther
Not set up for it?
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
llmrun Composite Score →
118.0
- 90% range 115.4–120.5
- #18 of 104 open models
- #16 of 101 that fit ≤640 GB
Run Mistral Large 4 1T A52B
One recommended way. Every other method, OS and quant is in the step-by-step guide.Each way opens a step-by-step guide for your OS and quant.
Recommended
vllm serve mistralai/Mistral-Large-4-1T-A52B --max-model-len 16384 --tensor-parallel-size 8Size the nodes from the table above
Download: Q4_K_M/Mistral-Large-4-1T-A52B-Q4_K_M-00001-of-00013.gguf634.80 GB + 12 partsggml-org
Ways to runOther ways
each opens step-by-stepRun in cloud
Rent 5× H200 SXM · Q4_K_M
- Estimated hourly price
- ~$17.95/hr
- est., whole server
- GPUs
- 5 GPUs
- 705 GB of VRAM
$5+ bonus for new users
How Much VRAM Does Mistral Large 4 1T A52B Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | Runs on | |
|---|---|---|---|---|
| 278.9 GB | 333.3 GB | 2-bit · i-quant extreme compression — Importance-weighted 2-bit, extreme compression — significant quality loss | Runs on 5× H100 SXM 80 GB | |
| 391.4 GB | 445.8 GB | 3-bit · i-quant — Importance-weighted 3-bit | Runs on 6× H100 SXM 80 GB | |
| 428.9 GB | 483.3 GB | 2-bit · K-quant — 2-bit quantization with K-quant improvements | Runs on 7× H100 SXM 80 GB | |
| 491.4 GB | 545.8 GB | 3-bit medium — 3-bit medium quantization | Runs on 8× H100 SXM 80 GB | |
| 603.9 GB | 658.3 GB | 4-bit medium — 4-bit medium quantization — most popular sweet spot | Runs on 5× H200 SXM 141 GB | |
| 716.4 GB | 770.8 GB | 5-bit medium — 5-bit medium quantization — good quality/size tradeoff | Runs on 6× H200 SXM 141 GB | |
| 828.9 GB | 883.3 GB | 6-bit, very good — 6-bit quantization, very good quality | Runs on 7× H200 SXM 141 GB | |
| 1003.9 GB | 1058.3 GB | 8-bit, near-lossless — 8-bit quantization, near-lossless | Runs on 6× B200 192 GB | |
| 2003.9 GB | 2058.3 GB | 16-bit · bfloat — Brain floating point 16 — preferred for training | Runs on 4 nodes of 8× H100 SXM |
Selected Q4_K_M
4-bit medium quantization629.3GBat 16K context
- Weights
- 600.0 GB
- KV cache
- 29.0 GB
- Overhead
- 0.3 GB
Server-class: ~629.3 GB at 16K context · about 5× H200 SXM 141 GB
Which Hardware Can Run Mistral Large 4 1T A52B?
Q4_K_M · 603.9 GBMistral Large 4 1T A52B (Q4_K_M) requires 603.9 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 786+ GB is recommended. Using the full 1049K context window can add up to 1851.8 GB, bringing total usage to 2455.7 GB. No consumer GPU has enough memory.
Needs server-class hardware
Mistral Large 4 1T A52B runs on a multi-GPU server · Q4_K_M needs ~629.3 GB
| Hardware | Quant | Memory | Speed | Context | Hardware page | |
|---|---|---|---|---|---|---|
| Mac 512 GB · Mac Studio (M3 Ultra, 512GB) | Q2_K | 454.3 GB | ~19 tok/s | 32K ctx | Mac Studio → | Q2_K · 454.3 GB · ~19 tok/s · 32K ctx |
| 8× H100 SXM 80 GB · 640 GB | Q3_K_M | 516.8 GB | — | 32K ctx | H100 SXM → | Q3_K_M · 516.8 GB · 32K ctx |
| 5× H200 SXM 141 GB · 705 GB | Q4_K_M | 629.3 GB | — | 16K ctx | H200 SXM → | Q4_K_M · 629.3 GB · 16K ctx |
| 4× B200 192 GB · 768 GB | Q4_K_M | 629.3 GB | — | 32K ctx | B200 → | Q4_K_M · 629.3 GB · 32K ctx |
Also: NVIDIA DGX A100 640GB · NVIDIA DGX H100
By VRAM tier · typical GPUs and Macs at Q4_K_M
Consumer GPUs (up to 32 GB) and Macs up to 256 GB: too small for this model (needs ≥ 304.3 GB)8 GB tier →12 GB tier →16 GB tier →24 GB tier →Macs →
How good is Mistral Large 4 1T A52B?
The llmrun Score merges 22 public benchmark boards into a single figure, where 100 represents the average for open models.
Score vs VRAM in the ≤640 GB tier
- Mistral Large 4 1T A52B
- Other models
- Best Score for its VRAM
#16 of 101 that fit ≤640 GB
Each dot is a model at the best quant that fits ≤640 GB. The line steps up to the best Score available at each amount of VRAM; a dot below it is beaten by a model that needs no more memory.
Full ≤640 GB ranking →Benchmarks
The 6 boards that weigh most in the llmrun Score, of 28 with a result for this model; badges show who ran the test.
- GPQA Diamond
Epoch AI · Oct 2026Independent
Independent88.1%
#16/100 open
- LiveBench Coding
LiveBench · Jun 2026Independent
Independent77.2
#8/20 open
- LiveBench Reasoning
LiveBench · Jun 2026Independent
Independent83.9
#9/20 open
- AIME 2024/2025
Epoch AI · Oct 2026Independent
Independent98.3%
#4/88 open
- FrontierMath
Epoch AI · Oct 2026Independent
Independent58.6%
#8/19 open
- LiveBench Math
LiveBench · Jun 2026Independent
Independent93.6
#2/20 open
In the ≤640 GB class
101 models fit ≤640 GBSame Score, only models that fit ≤640 GB at their best quant: this model's peers by memory need.
- #1
DeepSeek V4.1 Flash 1017DeepSeek
134.4
458.4 GB
- #2
MiMo V2.6 Pro RLXiaomi
133.3
614.1 GB
- #3
MiMo V2.6 Flash RLXiaomi
131.9
186.5 GB
- #4
GLM 5.3Z.ai
130.1
453.8 GB
- #5
DeepSeek V4.1 FlashDeepSeek
129.4
458.4 GB
- #6
Step 5 PreviewStepFun
128.0
407.5 GB
Alternatives to Mistral Large 4 1T A52B
Mistral family · 35 models →Official models with an llmrun Score, each answering one question; sizes are each model's default quant (usually Q4_K_M) with a 16K-token context.
Best alternative at this size
MiMo V2.6 Pro RL
Xiaomi
- 15.3 higher Score
- 15.2 GB less VRAM
Score 133.3 · 614.1 GB
Fits in less VRAM
DeepSeek V4.1 Flash 1017
DeepSeek
- 16.4 higher Score
- 170.9 GB less VRAM
Score 134.4 · 458.4 GB
Other vendor
GLM 5.3
Z.ai
- 12.1 higher Score
- 175.5 GB less VRAM
Score 130.1 · 453.8 GB
Similar models
10| Model | Score | VRAM | tok/s | Why it is similar |
|---|---|---|---|---|
| DeepSeek V4.1 FlashDeepSeek | 129.4 | 458.4 GB | —needs > 24 GB | similar score |
| DeepSeek V4 Pro 0813DeepSeek | 129.2 | 990.8 GB | —needs > 24 GB | similar score |
| Step 5 PreviewStepFun | 128.0 | 407.5 GB | —needs > 24 GB | similar score |
| Muse Spark Open 816B A42BMeta | 126.8 | 551.9 GB | —needs > 24 GB | similar score |
| GLM 5.2Z.ai | 122.9 | 453.8 GB | —needs > 24 GB | similar score |
| Qwen3.8 27BAlibaba | 121.9 | 17.9 GB | ~38 | similar score |
| Kimi K2.6Moonshot AI | 120.9 | 617.6 GB | —needs > 24 GB | similar score |
| Inkling SmallThinking Machines | 119.6 | 162.7 GB | —needs > 24 GB | similar score |
| Beam 1.0 501B A23Breflection | 119.5 | 304 GB | —needs > 24 GB | similar score |
| GLM 5.1Z.ai | 117.8 | 454.1 GB | —needs > 24 GB | similar score |
Specifications
- Publisher
- Mistral AI
- Family
- Mistral
- Parameters
- 1000B
- Active parameters
- 52B per token
- Architecture
- MistralLarge4ForConditionalGeneration
- Context Length
- 1,048,576 tokens
- Vocabulary Size
- 204,800
- Release Date
- 2026-09-30
- License
- Other
Frequently Asked Questions
How much VRAM does Mistral Large 4 1T A52B need?
Mistral Large 4 1T A52B requires 603.9 GB of VRAM at Q4_K_M, or 2003.9 GB at BF16. Full 1049K context adds up to 1851.8 GB (2455.7 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 1000B × 4.8 bits ÷ 8 = 600 GB
KV Cache + Overhead ≈ 3.9 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 ≈ 1855.7 GB (at full 1049K context)
VRAM usage by quantization
Can NVIDIA GeForce RTX 5090 run Mistral Large 4 1T A52B?
No — Mistral Large 4 1T A52B requires at least 278.9 GB at IQ2_XXS, which exceeds the NVIDIA GeForce RTX 5090's 32 GB of VRAM.
What's the best quantization for Mistral Large 4 1T A52B?
For Mistral Large 4 1T A52B, Q4_K_M (603.9 GB) offers the best balance of quality and VRAM usage. Q5_K_M (716.4 GB) provides better quality if you have the VRAM. The smallest option is IQ2_XXS at 278.9 GB.
VRAM requirement by quantization
★ Recommended — best balance of quality and VRAM usage.
Can I run Mistral Large 4 1T A52B on a Mac?
Yes, but only at lower quantizations. The smallest Mac that can run Mistral Large 4 1T A52B is Mac Studio (M3 Ultra, 512GB) at Q2_K; 1 of the 146 Macs we list can run it at some quantization. For Q4_K_M (603.9 GB) you need a Mac with more unified memory.
Can I run Mistral Large 4 1T A52B locally?
Yes — Mistral Large 4 1T A52B can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
What's the download size of Mistral Large 4 1T A52B?
At Q4_K_M, the download is about 600.00 GB. The full-precision BF16 version is 2000.00 GB. The smallest option (IQ2_XXS) is 275.00 GB.
Which GPUs can run Mistral Large 4 1T A52B?
No single consumer GPU has enough VRAM to run Mistral Large 4 1T A52B at Q4_K_M (603.9 GB). Multi-GPU or professional hardware is required.
Which devices can run Mistral Large 4 1T A52B?
Mistral Large 4 1T A52B requires at least 603.9 GB at Q4_K_M, which exceeds the usable memory of every unified-memory device we list (the largest has 508.5 GB).