Mistral AI announced Mistral Large 4 on October 6, 2026, describing it as an open-weight flagship for general agentic capabilities. The model is a mixture-of-experts system with about 1 trillion total parameters; Mistral reports about 49 billion active per token. It is available as a limited API preview, but its downloadable weights and license terms have not yet been released.
What Mistral Large 4 is
Mistral Large 4, nicknamed “le Chonk,” is Mistral AI’s new flagship model, announced October 6, 2026. The company positions it for general agentic capabilities: workloads in which a model can use tools or follow structured instructions as part of a task. Its public preview supports function calling, structured outputs, document Q&A, batching, and the Agents and Conversations endpoints.
| # | Preview | Product | Price | |
|---|---|---|---|---|
| 1 |
|
MINISFORUM MS-02 Ultra Workstation Mini PC, Intel Core Ultra 9 285HX (24C/24T, up to 5.5GHz), PCIe... | $1,659.00 | Buy on Amazon |
| 2 |
|
GMKtec EVO-X2 AI Mini PC Ryzen Al Max+ 395 Superchip 128GB LPDDR5X 2TB SSD | $3,649.99 | Buy on Amazon |
As an Amazon Associate I earn from qualifying purchases.
It is natively multimodal and is reported to support more than 160 languages, including every official European Union language. Secondary reports citing Mistral also give it a 1-million-token context window and a 1.6-billion-parameter vision encoder; those specifications have not been verified against a primary spec sheet, so treat them as reported rather than final.
Parameters, architecture and training
Mistral describes Large 4 as a mixture-of-experts (MoE) model. Its announcement puts the total at about 1 trillion parameters and the active count at about 49 billion per token. These figures describe different things: total parameters are the model’s overall scale, while active parameters are the portion used for a given token in an MoE model.
#1 Best Overall
- High-Performance AI Processor：The MS-02 Ultra features an Intel Core Ultra 9 285HX (24C/24T, up to 5.5 GHz, 13 TOPS NPU), delivering fast and efficient performance for AI inference, algorithm development, and media workloads. A PCIe x16 expansion slot supports desktop-class GPU upgrades for advanced model training and accelerated computing tasks. It's ideal for creators, engineers, and teams handling intensive parallel workloads.
- 4 × M.2 PCIe 4.0 + 4 × DDR5 SODIMM slots:Four DDR5 SODIMM slots support up to 256 GB of memory, while ECC helps maintain data integrity in mission-critical environments. Four PCIe 4.0 M.2 slots support up to 24 TB of storage, supporting RAID 0/1/5/10, combining high-speed performance with data protection. It allows for the creation of independent scratch disks, media libraries, and project drives, providing high-throughput for production workflows.
- PCIe & USB 4.0 v2: Up to three PCIe slots can be equipped, including a dual-slot x16 GPU. The main slot supports PCIe 5.0, meeting the needs of high-bandwidth creative and computing workloads. USB 4.0 v2 (80Gbps) supports high-bandwidth external storage and displays.
- Ultra-fast Networking: Wi-Fi 7 further enhances wireless performance with next-generation speeds and low-latency stability. Intelligent bandwidth switching optimizes throughput in different network environments, ensuring optimal performance for enterprise or local networks. Dual 25GbE ports (providing up to approximately 3.125 GB/s bandwidth, about 25 times faster than traditional 1GbE), enabling seamless large-scale file transfers and parallel computing. 10GbE and 2.5GbE ports, with support for Intel vPro technology, ensure enterprise-grade remote management and deployment flexibility.
- Server-grade thermal architecture: Utilizing a dedicated CPU/GPU airflow design, equipped with a 6-pipe dual-fan cooler, it maintains stable performance even under sustained loads, delivering up to 140W Turbo power while maintaining a 100W TDP, and operating with noise levels as low as 36 dB. An integrated 350W power supply ensures stable and reliable output for demanding computing tasks and fully loaded extended configurations.
There is a naming discrepancy to keep in view. The Hugging Face placeholder repository, “Mistral-Large-4.0-1T05-A52B,” suggests about 1.05 trillion total parameters and 52 billion active, rather than the announcement’s roughly 1 trillion and 49 billion. Until Mistral publishes final specifications, the safest reading is to attribute the 49-billion active figure to the announcement and note that the repository name points to a different approximate value.
Mistral says it trained the model from scratch over about two months in its European data centres, using about 4,000 Nvidia Grace Blackwell GPUs and roughly 10 megawatts of power. These are company-reported training figures, not a benchmark of the hardware needed to run the model.
Release date, API access and open weights
As of the announcement, the public route to Large 4 is a limited API preview in Mistral Studio. The model ID is mistral-large-4. The preview is not the same as a public weights release: the weights are still pending.
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Clear out junk files and repair common Windows errorsFree Scan →Scan for outdated or missing drivers - takes under a minuteDriver Scan →Mistral’s Hugging Face “Upcoming release” page lists October 31, 2026, for the weights. Some news reports instead give October 27; the official listing is the clearer date to follow, but it may change. Mistral has said the planned weight formats are FP8 and FP4. A gap between API preview and weights release is described in secondary reporting as allowing safety testing with trusted partners and governments.
The label “open-weight” should not be read as confirmation that Large 4 is open source. The weights are not yet available, and Mistral has not announced the license. Mistral Large 3’s Apache 2.0 license does not establish what terms will apply to Large 4. A secondary report has suggested a custom license, but that claim is unverified; wait for Mistral’s actual license before making deployment or redistribution plans.
API preview pricing
Secondary sources relaying Mistral’s pricing page report preview rates of $0.68 per million input tokens, $0.07 per million cached input tokens, and $2.09 per million output tokens. The same reporting showed crossed-out list rates of $1.36, $0.14, and $4.18 respectively, and described the lower rates as a roughly two-week promotion. Because the pricing and promotion are time-sensitive, check Mistral’s pricing page before estimating current costs.
Rank #2
- EVOLUTION RYZEN AI MAX+ 395 MINI PC - GMKtec EVO-X2 is the next evolution in AI mini PC Ryzen Strix Halo series. Thanks to AMD Simultaneous Multithreading (SMT) the core-count is effectively doubled, to 32 threads. Ryzen AI Max+ 395 has 64 MB of L3 cache and can boost up to 5.1 GHz, depending on the workload. The Ryzen AI Max+ 395 is currently rated as the "most powerful x86 APU" on the market for AI computing.
- AI NPU with XDNA 2 ARCHITECTURE - Powered by 16 “Zen 5” CPU cores, 50+ peak AI TOPS XDNA 2 NPU and a truly massive integrated GPU driven by 40 AMD RDNA 3.5 CUs, the Ryzen AI MAX+ 395 is a transformative upgrade and delivers a significant performance boost over the competition. The Ryzen AI Max+ 395 excels in consumer AI workloads like the llama.cpp-powered application: LM Studio. Shaping up to be the must-have app for client LLM workloads, LM Studio allows users to locally run the latest language model without any technical knowledge required and unleash their creativity and productivity.
- AMD RADEON 8090S iGPU GAMING PC - The AMD Radeon RX 8060S offers all 40 CUs with up to 2.9 GHz graphics clock and uses the new RDNA 3.5 architecture. The powerful iGPU is positioned between an RTX 4060 and 4070 laptop GPU and therefore enables gaming in FHD at maximum details in most demanding games. The 8060S can also utilize the full 128GB pool, which is perfect for running LLMs such as Deepseek 70B Q8, which runs comfortably on this machine.
- EIGHT CHANNEL LPDDR5X - LPDDR5X is a new ground breaking memory small form factor installed on-board. With blazing speeds up to to 8000MT/s, it runs 1.5x faster than the DDR5 SODIMMs; 90% better performance over DDR5 SODIMMs in video conferencing and photo editing; 30% better performance in productivity apps; 12% better performance in digital content workloads.
- QUAD SCREEN 8K DISPLAY SUPPORT - EVO-X2 AI Mini PC support 4-screen 4K/8K output via HDMI 2.1 (8K@60Hz), DisplayPort 1.4 (4K@60Hz), and dual USB 4 40Gbps Transfer speed (supporting PD3.0/DP1.4/DATA). Ideal for gaming, video editing, and multitasking, it provides expansive and crisp multi-display support.
For context, secondary coverage reported Large 3’s launch pricing as $0.50 per million input tokens and $1.50 per million output tokens. That historical comparison does not predict Large 4’s eventual standard rates.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
What Mistral claims about performance
Mistral says Large 4 is the best open-weights model from the United States or Europe on aggregated benchmarks, and that it outperforms closed frontier models on visual grounding. The benchmark results are the company’s own and are preliminary; Mistral says they may change as reinforcement learning is still under way. They should not be treated as independent confirmation of those rankings.
Mistral also acknowledges that the model still trails other frontier systems in coding. That matters if coding is the primary use case: the available information does not establish Large 4 as a leading coding model, even as Mistral makes broader claims about its benchmark standing.
Who can use which version
The public API preview and the planned downloadable weights are not the only access tiers. Mistral says developers, cybersecurity firms, and government agencies will receive a less restricted, more cyber-capable version than the public API. The public preview therefore should not be assumed to expose the model’s full capabilities or policy settings.
For teams considering self-hosting, Large 4’s eventual FP8 and FP4 weights, license, and hardware requirements will determine whether deployment is practical. A secondary headline claims the model fits on four datacenter GPUs, but the available details do not establish the hardware configuration, quantization, performance target, or deployment conditions behind that claim. It is not enough to size a production server.
Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Fix the driver behind crashes, sound loss and screen glitches3Clear out junk files and repair common Windows errorsWhat to compare before choosing it
- Openness: The model is called open-weight, but access to weights is pending and the license is unknown. Do not assume commercial-use, modification, or redistribution rights.
- API versus self-hosting: The Studio preview is the available option now. Self-hosting depends on the actual weights, license, and published hardware guidance.
- Performance evidence: Treat the current benchmark claims as preliminary and vendor-reported, particularly for coding comparisons.
- Multimodal and language fit: Mistral reports native multimodality and support for 160-plus languages; the 1-million-token context and vision-encoder details are secondary-source reports, not confirmed primary specifications.
- Data location and policy: Training in European data centres does not by itself establish where API requests are processed or stored. Check Mistral’s service terms and data-residency documentation for your deployment.
- Restricted access: Organizations in cybersecurity and government may have access to a less restricted tier that is not equivalent to the public API preview.
No specific, independently verified competitor figures are available here for a numerical comparison with DeepSeek, Qwen, Kimi, or other open models. Mistral’s broad positioning against frontier and Chinese open systems is not a substitute for matched, independent tests.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
