Mistral Large 4 Open Weights: Release Date, Specs and Hardware to Run It

Mistral AI put Mistral Large 4 into public preview on October 6, 2026, and promised that the open weights will follow by the end of the month. The Hugging Face page for mistralai/Mistral-Large-4-1T-A52B now shows a countdown to October 31, and nearly 900 people have clicked to be notified. It is a 1.05 trillion parameter, natively multimodal mixture-of-experts model with a 1M token context window, and Mistral calls it the best open weights model from the US or Europe. This guide pulls together everything confirmed so far: the release date, specs, benchmarks and pricing, what the license might look like, and an honest look at the hardware you would need to run Mistral Large 4 yourself, from a cloud server down to a home workstation.

Mistral Large 4 Open Weights Release Date

Here is the timeline as of October 22, 2026:

Mistral’s own wording is that the weights will be released “by the end of the month,” and the Hugging Face placeholder calls October 31 a current estimate, so a slip of a few days would not be surprising. I will update this post when the weights and license appear.

Specs: What We Know So Far

The “49B or 52B active” question confused a lot of people in the first week, since the launch post says one and the docs card says the other. The Hugging Face page settles it: 49B is the count for the transformer blocks, and 52B adds the embedding and output layers. That is also why the repo name ends in A52B.

Only about 4.7 percent of the weights are active for any given token. That matters a lot for running the model, as we will see below, because it means speed depends on the 49B active slice while memory depends on the full 1.05T.

Mistral Large 4 vs Mistral Large 3

Large 4 is about 55 percent bigger in total parameters but only about 20 percent bigger in active parameters, with four times the context. It is also nearly three times more expensive through the API, which tells you Mistral sees it as a frontier product rather than a cheap workhorse.

Benchmarks

All of these are vendor-reported, and the coding results were run privately, so treat them as claims until independent tests arrive with the weights.

The cybersecurity numbers are the headline. Mistral says several closed frontier models score near zero on the reproduce-and-patch test because they refuse the task, while Large 4 completes it. That, combined with self-hosting, is clearly the pitch to security teams: a capable model you can run on-premise, where your data never leaves the building.

In the Surge AI blind human evaluation of coding, Large 4 placed second of five models, behind Claude Opus 5 (4.22) and ahead of Kimi K3 (3.59). In finance and legal tasks, vals.ai reports it ahead of GPT-6 Astra on representative tasks, and Mistral claims it outperforms every open-source model on Harvey’s Legal Agent benchmark.

Pricing and How to Use It Today

You do not have to wait for the weights to try it.

The Ollama route is handy if you already use Ollama for local models, because the command is the same as for a local model:

ollama run mistral-large-4:cloud

Be clear about what that is, though. The :cloud tag sends your prompts to Ollama’s hosted servers. There is no downloadable local tag yet, and there cannot be until Mistral publishes the weights. Ollama also lists launch shortcuts for coding agents, such as ollama launch claude --model mistral-large-4:cloud and ollama launch opencode --model mistral-large-4:cloud, if you want to test it as a coding model.

Will Mistral Large 4 Be Apache 2.0?

Nobody outside Mistral knows yet. The launch post, docs card and Hugging Face placeholder all say “open weights” without naming a license.

The case for Apache 2.0 is precedent: Mistral Large 3 is on Hugging Face under Apache 2.0, and Mistral has leaned into “open” as a selling point in Europe. The case against is that Large 4 is a much more capable model with strong offensive security skills, and Mistral is red-teaming it with state authorities before release. A license with use restrictions, or a split between a research license and a commercial one, would not be a shock.

If you plan to build a product on Large 4, wait for the actual license file on release day before committing.

Hardware: What It Takes to Run Mistral Large 4

Mistral has not published hardware requirements, and nobody has shipped a quantized build yet, so everything in this section is arithmetic from the parameter count, not a measured result. The rule is simple: memory for weights equals parameters times bits per weight divided by eight. Real files run a little larger because of mixed-precision layers, and you need extra memory for the KV cache, which grows with context length.

Mistral Large 4: memory for the weights aloneEstimates from 1.05T parameters x bits per weight. KV cache for long context is extra.512 GB8x H200BF162,100 GBFP81,050 GBQ4-class (~4.5 bit)590 GBQ3-class (~3.5 bit)460 GBQ2-class (~2.6 bit)340 GBDynamic ~1.7 bit220 GBNot measured: official files are not out yet. Real quant files usually run a bit larger.

The FP8 guess is reasonable because Large 3 shipped in FP8, but it is a guess. For comparison, Mistral’s own card for Large 3 called for a single node of H200 or B200 GPUs. With roughly 55 percent more weights, Large 4 at FP8 fills a whole 8x H200 node before you add any context.

Why MoE makes home use possible at all

A dense 1 trillion parameter model would be hopeless outside a data center. With a mixture of experts, each token only touches about 49B parameters. At 4-bit that is roughly 28 GB of weights read per token. If those weights sit in system RAM with around 400 GB/s of bandwidth, the theoretical ceiling is somewhere around 10 to 14 tokens per second, and real-world numbers will be lower. That is slow but usable for batch jobs, coding questions and document analysis. It is the same logic that lets people run very large MoE models with llama.cpp’s expert offloading, where the shared layers live on the GPU and the experts stream from RAM.

Can You Run It at Home?

Here is a realistic breakdown by machine.

On my own RTX 5080 with 16 GB of VRAM, Large 4 is simply not a local model, and I would be suspicious of any guide that says otherwise. The 49B active slice alone is bigger than the card at any useful precision, and the full model is many times bigger than typical desktop RAM. That does not make it irrelevant to home users: the API is the practical way in, and open weights mean third-party hosts can compete on price, which tends to push costs down fast.

I covered the same question for Meta’s giant open model in Can You Run Meta Muse Spark Open Locally?, and the conclusions are similar: unified-memory machines and big-RAM servers are the realistic home path for trillion-scale MoE models.

What to Do on Release Day

1. Read the license first. Check commercial use, revenue thresholds and acceptable-use restrictions before you download a terabyte.
2. Check the official format. If the release is FP8 only, you will need a conversion step or a community quant for consumer setups.
3. Wait a few days for quants. The well-known GGUF and MLX quantizers usually publish within days of a big release, often with dynamic low-bit builds tuned for MoE models. Early quants of new architectures sometimes need llama.cpp updates before they load.
4. Check vLLM and SGLang support. For server deployments, confirm your inference engine has merged support for the new architecture, especially the vision encoder.
5. Start with short context. The KV cache for anything near 1M tokens is enormous. Test at 8K to 32K first.
6. Test your own tasks. Vendor benchmarks and 2-bit quality rarely match. Compare against the API on your real prompts.

FAQ

When is Mistral Large 4 open weights release?

Mistral promised the weights by the end of October 2026. The Hugging Face page for mistralai/Mistral-Large-4-1T-A52B shows an ETA of October 31, 2026.

How many parameters does Mistral Large 4 have?

1.05 trillion in total, with 49B active per token in the transformer blocks, or 52B including the embedding and output layers. It also has a 1.6B parameter vision encoder.

What is the context window?

1M tokens, according to Mistral’s docs. Some third-party hosts list less, for example OpenRouter showed 524K.

Can I run Mistral Large 4 with Ollama?

Only through the mistral-large-4:cloud tag, which runs on Ollama’s servers. A local download will only be possible after the weights are released, and it will need hundreds of gigabytes of memory.

Is Mistral Large 4 open source?

Mistral calls it open weights, but the license has not been published. Large 3 was Apache 2.0, and Large 4 may or may not follow.

Can an RTX 4090 or RTX 5090 run it?

Not on its own. Even a 2-bit build is roughly 340 GB. A high-end GPU helps only as an accelerator for the shared layers while the experts sit in a very large pool of system RAM.

Should I switch from Large 3?

If you self-host Large 3 on a single 8-GPU node today, Large 4 at FP8 will likely need more hardware. Benchmark both on your tasks through the API before planning a migration.

Final Thoughts

Mistral Large 4 is the biggest open weights release from a European lab so far, and if the license is permissive it could become the default self-hosted model for companies that cannot send data to US clouds. For individual users, the honest answer is that you will use it through an API long before you run it at home, and you need a 512 GB class machine to run it locally even at low precision. I will update this guide when the weights, license and first quantized builds land at the end of October.

Related guides on this site:

• Can You Run Meta Muse Spark Open Locally? Hardware, Quants and Setup
• Comfy Agent Guide: Pricing, Credits, Local Support and Comfy MCP
• FLUX 3 Image Klein in ComfyUI, an open model that does fit on a gaming GPU.
• Best Local AI Video Models in 2026

Sources

• Mistral AI: Introducing Mistral Large 4
• Mistral Docs: Mistral Large 4 model card
• Hugging Face: mistralai/Mistral-Large-4-1T-A52B (upcoming release)
• Ollama: mistral-large-4
• MarkTechPost: Mistral Large 4 specs and benchmarks
• System Design Academy: Mistral Large 4 release, weights and hardware