Search the full wire by company, model, lab, or keyword. Every story we have ever aggregated.
DeepSeek plans to train V5 model on Huawei Ascend 950 chips in 2027, extending DeepSeek-Huawei partnership from inference to full training.
Weight Space Programming enables LLMs to directly edit neural network weights; demonstrated on Qwen and DeepSeek for reasoning optimization.
DeepSeek previews autonomous 'super agent' for multi-task workflows with minimal human oversight, competing with OpenAI and Anthropic on coding agents.
, the open-source AI penetration testing tool, can now run on any of the 400+ models on Vercel .StrixAI Gateway Strix can reach models from Anthropic, OpenAI, Kimi, GLM, Deepseek, and others through a single API with no markup, removing the need to sign up for multiple providers. Usage and cost are tracked across all models in AI Gateway. Strix selects the Gateway through , where the model is the standard AI Gateway ID behind a prefix:STRIX_LLM / vercel_ai_gateway/ See to configure it and get started today.Strix's AI Gateway setup guide Read more
from Particle AI is now available on . It is post-trained on DeepSeek V4.1 Flash, with a focus on completing software engineering tasks using fewer tokens. Particle describes the gains as shorter reasoning, fewer rounds of trial and repair, and more focused exploration.Particle-1AI Gateway Use it for debugging, planning code changes, and breaking down multi-step engineering problems. Particle-1 has a 1M-token context window for working with code and extended task histories. Set in your environment. Use as the model name:AI_GATEWAY_API_KEYparticle/particle-1 To use it in Claude Code, Codex, Cu...
Independent third-party safety evaluation of DeepSeek V4.1 Flash-1017 for loss-of-control and manipulation risks following EU AI Code of Practice.
Debate over whether China's open-weight models from DeepSeek, Z.ai, and Alibaba pose cybersecurity risks or enhance internet security.
Proposes low-precision storage for n-gram embedding tables in language models, targeting 196B-parameter tables in DeepSeek-V4.1-Flash without performance loss.
DeepSeek has reportedly raised $15 billion in one of China's largest funding rounds, just days after the US ordered federal agencies to remove its technology over national security concerns.
DeepSeek releases DeepInfer, open-source inference stack processing 155T tokens/day at ~80% margin, enabling $22M/day revenue with 81-day payback period.
The US has ordered federal civilian agencies to remove DeepSeek products and third-party services built using its technology, citing national security concerns.
DeepSeek V4.1 Flash 1017, MiniMax M3.1 Flash, and Mercury 3 released; limited immediate adoption expected.
now includes DeepSeek V4.1 Flash, alongside Kimi K3 and GLM 5.3, each accessed through its preferred provider, with more models to come. The DeepSeek model uses the latest weights (1017).Token Pass Token Pass is a $30/month subscription on Pro teams that grants $80 in monthly credit and runs through . Use it with any coding agent such as , , , etc. or build your own with the AI SDK, with Zero Data Retention enabled by default.AI GatewayfxPiClineCommand Code Add DeepSeek V4.1 Flash from the in your team's AI Gateway dashboard, or read the for setup.Token Pass pageToken Pass documentation Read ...
is now available on .DeepSeek V4.1 Flash 1017AI Gateway DeepSeek V4.1 Flash 1017 is a minor update to that improves reasoning, coding, agentic tasks, and image understanding.DeepSeek V4.1 Flash The coding gains show up in terminal work and in building and editing whole repositories, and the agentic gains in multi-step automation. With images, the model reads charts and handles visual reasoning more accurately when it can call tools. DeepSeek V4.1 Flash 1017 also finishes tasks faster and with shorter outputs than the previous snapshot, so agent runs tend to take less time and use fewer output...
Controlled study of knowledge injection into addressed memory on GPT-2, Qwen3.8-Flash, and DeepSeek-V4.1-Flash examining transfer and behavior preservation.
Evaluation of DeepSeek-V4-Flash and GLM-5.3-Flash on LLM entity extraction from cyber threat intelligence reports shows failure modes and confidence calibration patterns.
Comparison of LLM retrieval errors in environmental science journals: Claude, ChatGPT, Grok, DeepSeek, Perplexity, others tested on 2024-2025 citations.
ID Balancing applies PID control to stabilize training of extremely sparse MoE models, unifying DeepSeek's loss-free method with proportional-integral-derivative control.
Study analyzes reverse reasoning ability in GPT-o1, GPT-o3, and DeepSeek-R1, identifying five error categories in complex problem-solving.
Mixture of experts (MoE) has become one of the defining architectural trends in large-scale AI model training. DeepSeek, Qwen, and Mixtral are examples of MoE... Mixture of experts (MoE) has become one of the defining architectural trends in large-scale AI model training. DeepSeek, Qwen, and Mixtral are examples of MoE models that match or exceed the performance of dense model counterparts at a fraction of the training compute. MoE models provide efficient training through conditional computation. Instead of one dense feed-forward network (FFN) shared… Source
DeepSeek releases v4.1-Flash with 763B parameters, novel causal encoder-decoder architecture, and vision capabilities.
A new report released Thursday by Anthropic alleges persistent distillation attacks by China-based AI companies, which have escalated in recent months as competition in the space has intensified.
Multi-model LLM evaluation for educational short-answer scoring using GPT, DeepSeek, Qianwen across reliability, validity, and failure modes.
Mechanistic analysis of DeepSeek-V4-Flash's four-stream residual pathways reveals selective routing and distinct representation streams.
LLM-labeled NER training for Persian customer chat anonymization, comparing DeepSeek-V3, Qwen3, and GPT-OSS labeling quality.
Qwen 3.8 27B scores 52 on the Artificial Analysis Intelligence Index That's the same score as GPT-5.6 Luna (max), and just one point behind GLM-5.2 (max) and DeepSeek V4 Pro 0813 (max) - that GLM is 753B and that DeepSeek is 1.6B parameters, and Luna is size unknown but presumably a whole lot bigger than 27B. Qwen 3.8 27B is a truly astonishing model . Via Hacker News Tags: ai , generative-ai , llms , qwen , ai-in-china , artificial-analysis
DeepSeek V4 Pro 0813 (on OpenRouter) The latest DeepSeek Pro model is now available, via API only. I had to link to OpenRouter because DeepSeek don't have any obvious announcement page for their new model. I haven't been able to confirm if they plan to release the open weights, but given the weights are available for both April's deepseek-ai/DeepSeek-V4-Pro and July's deepseek-ai/DeepSeek-V4-Flash-0731 it seems likely. Interestingly I got very different looking pelicans for the three different reasoning levels of low, medium, and high. I've not noticed this kind of difference from any other m...
Simon Willison's June 2026 newsletter roundup covering model releases (GPT-5.6, Claude Opus 5, DeepSeek-V4), open letters, and accidental cyberattacks by OpenAI and Anthropic test models.
DeepSeek releases V4-Flash-0731, a 304B parameter model with enhanced agentic capabilities, outperforming larger competitors at $0.14/$0.27 per million tokens.