TokenPost topic
Mistral AI
13 TokenPost stories are tagged Mistral AI. The newest appear first.
Latest Mistral AI stories
- Mistral Brings Vibe Work and Coding Tools to a Desktop App
Vibe Desktop became generally available Tuesday, Nov. 17, for macOS and Windows, with coding tasks available on local repositories or in managed cloud sandboxes.
· Technology
- Mistral Says Mistral Large 4 Coordinated 133 Agents in Django Security Demo
The video ran at 90 times normal speed. Django had published fixes for six issues on Nov. 3, but Mistral did not identify the tested flaw or show its patch.
· Technology
- Banque de France Plans Mistral AI Rollout on a Sovereign Cloud
The projects are part of a doctrine guiding the bank’s 2028 strategic roadmap, which emphasizes control of data, infrastructure and operating skills.
· Technology
- X Post Credits Mistral Large 4 With a Satellite Tracker, but Gives Few Build Details
The post describes Orbital Trace as a live satellite and debris tracker but does not explain the model’s role or where to access the software.
· Technology
- Mistral Large 4 Demo Pairs Hartree-Fock Solver With 3D Viewer
Users can stretch benzene or water bonds while a browser page recalculates molecular orbitals. The generated solver checked its results against an unnamed chemistry tool.
· Technology
- Mistral Opens Responses API Preview for Developers Building AI Agents
The API returns answers, reasoning and tool calls as separate items. It supports Mistral’s hosted chat models, including Large 4.
· Technology
- Mistral’s Voxtral Mini Ranks Second by Diarization Error Rate
The four-billion-parameter model posted a 40.4% diarization error rate, ahead of Voxtral Small Transcribe 3 at 41.1%, while Small had a lower AA-WER.
· Technology
- Mistral AI, NVIDIA Expand Partnership Into Engineering Simulation
The companies plan public datasets and benchmarks for physics models, alongside tests of AI agents performing engineering tasks.
· Technology
- Mistral Launches Voxtral 3 Voice-Agent API Starting at $0.04 Per Minute
The starting rate uses Mistral Small 4 for reasoning. The API costs $0.08 per minute with another reasoning model.
· Technology
- Accenture CTO Says Few Large Enterprise or Government Processes Center on AI
Under its 2027–2029 strategy, AXA expects AI to generate €500 million to €700 million in recurring annual pre-tax benefits by 2029. In June 2026, the company said 400 AI use cases were in the scaling phase.
· Business
- CERN and Mistral Plan AI Model for Particle Physics Research
The Large Physics Model would train on billions of simulated LHC collisions and support reconstruction, particle identification, signal detection and detector design.
· Technology
- Nvidia Details European AI Infrastructure Expansion Across Four Countries
The company is deploying Blackwell computing capacity through government, technology and cloud partnerships, while individual investment amounts remain undisclosed.
· Technology
- Mistral AI Releases Large 4 With One Trillion Parameters
The model was trained using 4,000 NVIDIA GPUs. Mistral plans to release its weights after safety testing in about three weeks.
· Technology