Models & Labs
DeepSeek V4 Paper Released
AI Explainedmedium confidence
Why it matters
- →DeepSeek V4's advancements could set new standards in AI data processing tools.
DeepSeek has released its V4 paper, detailing the latest advancements in its AI capabilities. This version introduces new features aimed at enhancing performance and efficiency in data processing tasks. The paper includes comparisons with previous versions and highlights the improvements made. This release is expected to influence the competitive landscape in AI tools and applications.
Read originalThe story around this
Earlier coverage that leads up to this article, and what followed. Lines connect each piece to the closest one after it, converging here.
What happened next
- 1DeepSeek-V4 Pro Offers Permanent 75% Discount
Lev Selector · May 29, 2026 · Related
- 2DeepSeek V4 Flash Released
Lev Selector · July 31, 2026 · Same story
- 3DeepSeek-V4-Pro Released for Enhanced Search
Matt Wolfe · August 14, 2026 · Same story
- 4DeepSeek V4 Pro 0813 vs Claude Fable 5: Cost and Efficiency
Together AI Blog · August 17, 2026 · Related
- 5DeepSeek V4 Pro 0813 vs GPT-5.6 Sol: Cost Efficiency
Together AI Blog · August 18, 2026 · Related
Follow this story
Open the full story →DeepSeek V4 Paper Released
10 developments
- Apr 24 · AI Explained
DeepSeek V4 Paper Released (This article) - Apr 24 · Lev Selector
DeepSeek V4 Launches with 1.6 Trillion Parameters↳ DeepSeek V4 has 1.6 trillion parameters. - Apr 24 · MIT Technology Review AI
DeepSeek launches new open-source AI model V4↳ DeepSeek V4 can process longer prompts than previously stated. - Apr 27 · The Rundown AI
DeepSeek Launches Affordable V4 AI Model - Apr 28 · The AI Daily Brief
DeepSeek V4 Launches with Million-Token Context - Apr 29 · Together AI Blog
DeepSeek-V4 Pro Launches on Together AI↳ DeepSeek-V4 Pro launches on Together AI with 512K context, controllable reasoning modes, and cached-input pricing. - May 1 · Matt Wolfe
DeepSeek V4 Preview Released - May 1 · Lev Selector
DeepSeek V4 Pro Launches with 1.6T Parameters - May 2 · Matt Wolfe
DeepSeek V4 Offers Cost-Effective AI Solution - May 8 · Together AI Blog
DeepSeek-V4 Tackles Million-Token Context Challenge↳ DeepSeek-V4 uses hybrid attention to compress context before KV storage, reducing memory pressure on NVIDIA HGX B200.
More from AI Explained
© AI ExplainedResearchresearch
New Benchmarks: ExploitGym, BenchCAD, and IntegrityBench
Several new evaluation benchmarks have been released, including ExploitGym for security vulnerabilities and BenchCAD for design robustness.
AI Explained
© AI ExplainedResearchresearch
METR Report: AI R&D Acceleration at Anthropic
A new METR report assesses the acceleration of AI research and development within Anthropic, analyzing internal model improvements.
AI Explained
More in Models & Labs
Models & Labsother
llama.cpp b12340 adds CUDA 13 and ROCm 10.0
This release quietly closes the hardware gap for local inference by adding default support for CUDA 13 and ROCm 10.0 across Linux and Windows. Users with newer NVIDIA GPUs or AMD accelerators no longer need to compile from source to get performance, while Apple Silicon builds now ship KleidiAI disabled by default, shifting the focus back to standard ARM optimizations. The inclusion of Snapdragon support for Linux and Android marks a significant step toward on-device AI beyond the x86/ARM desktop ecosystem. It is less about new features and more about ensuring llama.cpp remains the universal runtime for every available accelerator.
llama.cpp Releases
Models & Labscoding
llama.cpp b12342 adds few-row MMQ for speculative decoding
Speculative decoding just got a significant speed boost on NVIDIA GPUs. This release introduces a specialized kernel for 'few-row' matrix multiplications, targeting the small batch sizes typical of verification passes in speculative decoding workflows. By optimizing these specific operations for quantized formats like Q8_0 and IQ4 variants, llama.cpp reduces latency where it matters most during token generation. The change is CUDA-only for now but leverages hardware capabilities from Ampere onwards, making local inference faster without requiring new model architectures or complex configuration changes.
llama.cpp Releases
© TechCrunch AIModels & Labsmodels
Google deploys Gemini 4 Hydrogen for faster Search
Google is quietly shifting its search infrastructure by replacing older fast models with Gemini 4 Hydrogen, a tier designed specifically for speed and cost-efficiency rather than raw reasoning power. This model completes tasks 2.5 times faster than its predecessor while beating the more powerful Argon on specific intelligence benchmarks, proving that specialized efficiency can outperform brute force in high-volume workflows. By integrating this into Search's fan-out technique, Google now breaks down complex queries into subqueries with greater precision and speed, delivering richer visual responses without the latency of heavier models. This marks a strategic pivot toward agentic execution layers within consumer search, making AI Mode feel snappier for everyday multi-part questions.
TechCrunch AI