DeepSeek has released DeepSeek-V4-Pro-Base, a 1.6 trillion parameter foundation model now available on Hugging Face. The base model uses BF16 precision for weights and includes support for F8_E4M3, I64, and F32 tensor types.
DeepSeek has released V4 Pro, a Mixture-of-Experts model with 1.6 trillion total parameters and 49 billion activated parameters. The model supports a 1-million-token context window and costs $1.74 per million input tokens and $3.48 per million output tokens.
DeepSeek has released V4 Flash, a Mixture-of-Experts model with 284B total parameters and 13B activated parameters per request. The model supports a 1,048,576-token context window and is priced at $0.14 per million input tokens and $0.28 per million output tokens.
Alibaba announced Friday that its Qwen AI model will be integrated into vehicles from 10 Chinese automakers including BYD, Geely, Li Auto, and SAIC Volkswagen. The system runs on Nvidia's automotive chip platform and allows drivers to order food delivery, book hotels, and make payments through voice commands, even with limited network connectivity.
DeepSeek has released two new models in its V4 series: DeepSeek-V4-Flash and DeepSeek-V4-Pro, both now available on Hugging Face. Separately, Tencent has shipped Hy3-Preview, marking simultaneous releases from two major Chinese AI labs.
DeepSeek released two Mixture-of-Experts models: V4-Flash with 284B total parameters (13B activated) and V4-Pro with 1.6T parameters (49B activated). Both models support one million token context windows and use a hybrid attention architecture that requires only 27% of the inference FLOPs compared to DeepSeek-V3.2 at 1M token context.
DeepSeek released two new Mixture-of-Experts models: DeepSeek-V4-Pro with 1.6 trillion parameters (49B activated) and DeepSeek-V4-Flash with 284B parameters (13B activated), both supporting one million token context length. The models achieve 27% of inference FLOPs and 10% of KV cache compared to DeepSeek-V3.2 at 1M context through a hybrid attention architecture combining Compressed Sparse Attention and Heavily Compressed Attention.
Anthropic confirmed that widespread complaints about Claude Code quality degradation were caused by three separate bugs in the coding assistant's harness, not the underlying models. One critical bug caused Claude to clear its thinking context every turn in sessions that had been idle for over an hour, making it appear forgetful and repetitive.
Anthropic confirmed three separate system changes in March and April degraded Claude Code, Claude Agent SDK, and Claude Cowork performance. The company reduced default reasoning effort from high to medium on March 4, introduced a caching bug on March 26 that cleared session data with every turn, and added restrictive word limits on April 16 that caused a 3% performance drop.
Anthropic has expanded Claude's integration directory to include 15 lifestyle services including Spotify, Instacart, AllTrails, Uber, and Booking.com. The update shifts Claude's third-party connectivity from professional and educational tools to personal use cases, with apps now appearing dynamically within conversations.
Anthropic has released app connectors allowing Claude to integrate directly with personal services including Spotify, Uber, Uber Eats, Instacart, TurboTax, Audible, AllTrails, and TripAdvisor. The connectors are available now across all Claude plans, with mobile support in beta.
Anthropic has released version 0.97.0 of its Python SDK, introducing CMA Memory as a public beta feature. The update includes bug fixes for API spec errors, restored missing features, and performance improvements for multipart file requests.
Anthropic's Claude Opus 4.7 release triggered a sharp increase in false positive refusals, with developers filing 30+ complaints in April 2026 compared to 2-3 monthly reports from July-September 2025. The company deployed aggressive Acceptable Use Policy guardrails to prepare for the eventual release of its Mythos vulnerability research model.
Tencent has released Hy3-preview, a 295-billion-parameter Mixture-of-Experts model with 21 billion active parameters and a 256K context window. The model scores 76.28% on MATH and 34.86% on LiveCodeBench-v6, with particularly strong performance on coding agent tasks.
OpenAI released GPT-5.5, which completes tasks at the same difficulty level faster than GPT-5.4 while using significantly fewer tokens. The model is available to ChatGPT Plus, Pro, Business, and Enterprise users, with API access delayed due to required safeguards.
OpenAI has released GPT-5.5, its latest frontier model, according to NVIDIA. The model powers Codex, OpenAI's agentic coding application, running on NVIDIA GB200 NVL72 rack-scale systems.
OpenAI released GPT-5.5 on April 23, 2026, seven weeks after GPT-5.4. The model features a 400K context window in Codex and claims improvements in multi-step tasks, agentic coding, and computer use, though at higher pricing than its predecessor.
OpenAI released GPT-5.5 on April 23, 2026, positioning it as a step toward agentic computing and a unified 'superapp' combining ChatGPT, Codex, and browser capabilities. The company claims the model outperforms GPT-5.4, Google's Gemini 3.1 Pro, and Anthropic's Claude Opus 4.5 across multiple benchmarks.
OpenAI has released GPT-5.5, one month after GPT-5.4, claiming improved performance on coding tasks and reduced token usage in its Codex platform. The model rolls out April 24 to paid ChatGPT tiers.
OpenAI released GPT-5.5, its latest AI model with enhanced coding, computer operation, and research capabilities. The model is rolling out to paid subscribers in ChatGPT and Codex, with API access coming soon.