Mistral Large 4
AI overview
Sign in and the AI will write an overview from our coverage.
Headlines · 19
- vLLM v0.34.0 adds 10 models and enables extensible KV cache by default
vLLM v0.34.0 introduces 10 models, including Mistral Large 4, MiniMax-M3.1-Flash and NVIDIA Nemotron-3.5-Transcribe, and makes extensible KV caching the default. The release removes Model Runner V1, requires huggingface_hub 2.0.0 or later, changes the default block size to 64 and drops ROCm 7.2 artifacts. vLLM reports up to 1.95× faster decoding, 1.6× higher throughput and 1.24–1.40× faster long-prompt TTFT on Blackwell and MI355X hardware.
vLLM Releases · 🔥 0 - Deepnote Benchmark Ranks AI Models for Proactive Workplace Help
Deepnote’s Agent Initiative Benchmark tested 12 models in a simulated company Slack environment, asking them to determine when to respond, assist privately, or remain silent across 50 workplace tasks. GPT-6 Astra scored 8.97 and GPT-6.1 Sol scored 8.65, while Claude Fable 5.2 scored 8.17 at a per-task cost reported as 17 times Sol’s; the dataset and code were released openly. The results also showed that models often responded when they should have stayed silent, with unnecessary interventions consuming more time and money.
新智元 · 🔥 0 - Claude Opus 5.5 leads OpenRouter scanned-document OCR test
OpenRouter tested 15 vision models, Mistral OCR and Tesseract on eight synthetic pages designed to resemble scanned receipts, letters, forms, newsletters, handwriting and faxes. Claude Opus 5.5 recorded the highest key-value accuracy, getting 262 of 264 checks right, while Mistral OCR got 258 of 264 at one-eighth the cost. The report also found that vision models can fabricate values, while Tesseract struggled with faded print, handwriting and faxes. OpenRouter said the small synthetic test set is only a starting point and recommended testing candidates on users’ own scans.
OpenRouter Blog · 🔥 0 - Mistral Large 4 agents test Django vulnerability triage and patching
Mistral AI says 133 agents powered by Mistral Large 4 mapped code, searched for flaws, tested attack scenarios and confirmed a vulnerability in a recent Django CVE. After reproducing the issue, the vendor says its Le Chonk system patched it, demonstrating a claimed triage, reproduction and patching workflow.
Mistral AI (X) · 🔥 0 - Mistral Large 4 joins Open Math OS
The SAIR Foundation announced that Mistral Large 4 is now available to approved teams using the Open Math OS research workspace. It described the model as having “1.05T open weights” and a 1-million-token context window. Open Math OS is a workspace where mathematicians and AI agents collaborate on shared projects using open models that the community can inspect and run, with community control of project data.
Mistral AI (X) · 🔥 0 - Mistral Large 4 Builds Orbital Trace 3D Satellite Tracker
Mistral AI says it tasked Mistral Large 4 with building Orbital Trace, a real-time 3D tracker for roughly 19,000 satellites and debris fragments orbiting Earth. According to the company, the model built a working app, tested it as a user, optimized its speed and passed strict verification.
Mistral AI (X) · 🔥 0 - Mistral Large 4 Demonstrates AI-Generated Quantum Chemistry Workflow
Mistral AI says Mistral Large 4 wrote a quantum chemistry self-consistent field (SCF) solver from a short prompt, validated its output and built a 3D molecular view. The company also says the model chose how to break the molecule and which orbitals to show without further input. The supplied report does not include an independent reproduction or technical evaluation.
Mistral AI (X) · 🔥 0 - Mistral Launches Large 4 for Enterprise Cybersecurity
Mistral AI released Large 4, abbreviated ML4, with a reported total parameter count of 1.05 trillion and roughly 52 billion active parameters per task. The company says the model offers advanced programming and cyber-defense capabilities and can run on enterprises' own computers and servers; 36Kr reported that it was trained in Mistral's European data center using 3,800 Nvidia Grace Blackwell GPUs.
36氪快讯 · 🔥 0 - Orange Cyberdefense joins Mistral Large 4 design partner programme
Orange Cyberdefense has joined the design partner programme for Mistral Large 4, becoming the first European managed security services provider to participate. It plans a gradual rollout across its services, with initial uses including technical audits, penetration testing, analysis of how attackers might gain access to systems and incident investigations. The cybersecurity-focused model will run on computers in Mistral AI’s European data centres.
The Next Web · 🔥 0 - Mistral Releases Mistral Large 4 Open Weights
Mistral AI released the open weights for Mistral Large 4, also known as Le Chonk. It is a native multimodal, multilingual mixture-of-experts model with about 1 trillion total parameters, 49 billion active parameters and a 1-million-token context window. Official FP8 and NVFP4 variants are available on Hugging Face, alongside community-created GGUFs from Unsloth and the llama.cpp team. The model is offered under the Mistral Ecosystem Open-Weight License, which permits use, modification, fine-tuning and commercial use, while requiring a separate commercial license for model-as-a-service businesses. Mistral says the model leads open-weight coding and defensive-security benchmarks and offers three reasoning modes. Reported external results include 19.7% on Cognition's private FrontierCode 2.0 benchmark and 48% on Artificial Analysis's Coding Agent Index, matching GLM-5.3 while using 47% fewer tokens per task on average.
Mistral AI News · 🔥 0 - vLLM adds day-0 support for Mistral Large 4
vLLM announced day-0 support for Mistral Large 4, Mistral’s native multimodal mixture-of-experts model, through the dedicated vllm/vllm-openai:mistral-large-4 Docker image. The implementation supports FP8 and NVFP4 checkpoints, image inputs, built-in DSpark speculative decoding, long-context serving and AMD Instinct MI350 deployment with ROCm. It also includes generic improvements to context parallelism, NIXL-based disaggregated serving and HiSparse host-memory KV offload. Performance and quality figures reported by vLLM were not independently verified in the supplied material.
vLLM Blog · 🔥 0 - 12 AI Models Share a Conformity-Themed View of Humans
R1’s test team asked 12 text models from 10 laboratories in three countries for their most controversial, unhedged take on humans three times. Every model gave a conformity-themed answer at least once, saying people prioritize group beliefs, belonging, status or personal stories over truth or principles; seven of the 12 gave that shared theme in all three answers. The study used no system prompt or retries and made 60 calls costing $1.54.
Hacker News · AI(100+ 分) · 🔥 0 - Mistral and NVIDIA plan physics AI, engineering-agent benchmarks
Mistral announced on October 21 a collaboration with NVIDIA to build AI-native engineering workflows that combine Mistral Large 4 and Mistral's data-generation and physics-modeling pipelines with NVIDIA CUDA-X, Warp and Omniverse libraries, aiming to speed simulation and connect design, simulation and model training in closed loops. The partners also said they plan to publish open datasets and public benchmarks for evaluating physics models, plus dedicated benchmarks for agentic AI in computer-aided engineering and electronic design automation that measure tool use, multistep workflow completion, technical correctness, execution time and computational cost. Mistral describes Mistral Large 4, nicknamed Le Chonk, as a multimodal model with roughly 1 trillion total parameters and 49 billion active parameters, trained on 3,800 NVIDIA Grace Blackwell GPUs in Europe, with open weights planned for release at the end of October; these details come from Mistral and were repeated by The Decoder.
Mistral AI News · 🔥 0 - Meta releases Muse Spark Open, the strongest US open-weight model, analyst says
Interconnects argues that Meta's Muse Spark Open, released this week, is now the strongest American-made open-weight model by a wide margin. According to the Hugging Face model card cited in the piece, it has 816B total and 42B active parameters, a 1M-token context window, native multimodality and is built for coding and agentic work. The analysis places its performance above GLM 5.2 and near DeepSeek V4.1 Flash, but behind GLM 5.3 and Kimi K3, and notes that Meta also unveiled the closed frontier models Muse Wonder and Muse Wonder Full days earlier. It says the gap between the best US and Chinese open models has been roughly halved.
Interconnects · 🔥 0 - Mistral Large 4 preview tops non-Chinese, non-US models on Intelligence Index
The Batch reports that Mistral trained Mistral Large 4 on 3,800 GPUs in its own data centers, using data spanning more than 160 languages, and mixed many task types — single-turn chat, science problems, safety, factual accuracy and long-running tool use — into one reinforcement learning run with rewards from unit tests, reward models, LLM judges and code inspection without execution. On Artificial Analysis' Intelligence Index, the API preview in reasoning mode scores 38 at $1.13 per task, tying GPT-6 Luna and trailing MiMo-V2.6-Pro (46), GLM-5.3-Flash (42) and DeepSeek V4.1 Flash (39). On Artificial Analysis' Cyber Index, which measures finding and fixing vulnerabilities in code repositories, the preview reaches 50 percent, tying GLM-5.3-Flash. The Batch notes the RL run that produced the preview is still in process.
The Batch · DeepLearning.AI · 🔥 0 - Mistral releases trillion-parameter Large 4, the best model outside the US and China
European AI lab Mistral has released Large 4 (also called ML4 and nicknamed "Le Chonk" on social media), a trillion-parameter model. Testing lab Artificial Analysis scored it 38 on its overall intelligence index, according to Tom's Hardware, leaving it behind Chinese models such as Kimi K3 (max), GLM-5.3-Flash and MiMo-V2.6.Pro, though it performs strongly on cybersecurity tasks. Tom's Hardware reported that independent tests rank ML4 as the best AI model outside the U.S. and China, and said Mistral expects the exact weights to be released before the end of October.
Tom’s Hardware · 🔥 0 - Anthropic, Google, Mistral roll out Haiku 5.5, Gemini 4 Argon and more
Over the past week Anthropic, Google and Mistral AI each shipped new models. Anthropic released Claude Haiku 5.5 — its first major refresh in a year for its cheapest, fastest Claude — which it says costs about 75% less to run than the previous Haiku and is the first Haiku-level model with an adjustable effort setting, alongside Claude Sonnet 5.5, which posts sizable gains in agentic coding and knowledge work and alignment and cybersecurity capabilities the company compares to Opus 5. Google's Gemini 4 Argon, a frontier model with a 1 million-token output limit and stronger deep reasoning, is limited to trusted cybersecurity workers through its Fairwind program, while Mistral's flagship Mistral Large 4, nicknamed "Le Chonk," is in public preview and, per Mistral, ranks among the world's top open-weight models ahead of an Oct. 27 rollout.
CNET · 🔥 0 - llm-mistral 0.16 adds support for reasoning models
Simon Willison released llm-mistral 0.16, which adds support for reasoning models to the LLM command-line tool. The post cites the newly released Mistral Large 4 as an example. No further details about the update or the model are given in the source.
Simon Willison's Weblog · 🔥 0 - Mistral launches Mistral Large 4 preview, a 1T-parameter open-weight model
Mistral AI has opened a public preview of Mistral Large 4 (nicknamed Le Chonk), a 1-trillion-parameter natively multimodal open-weight model with roughly 49 billion active parameters; the weights are due by the end of the month. Mistral claims state-of-the-art open-model performance on enterprise workloads such as cybersecurity, finance and law, and says it beats even closed frontier models on visual grounding — though it notes that some cyber-test leads come from Claude and GPT-6 refusing the tasks. Independent measurement is more sobering: Artificial Analysis's Intelligence Index puts ML4 at 38 points, a big jump from Mistral Large 3's 9, but well behind leader Claude Opus 5.5 (Max) at 58.
Mistral AI News · 🔥 0
Experience and discussion from the community
Share my Mistral Large 4 experienceAsk about Mistral Large 4
Nobody has shared their experience with Mistral Large 4 yet.