Developer intelligence
New models, model updates, capabilities, benchmarks, pricing, and availability.
Your stack
Add your stack to personalize.
sglang v0.5.21 introduces support for top open LLMs and diffusion models, major API updates for low-latency scoring/classification, performance boosts for popular models, and a new Rust prefix cache.
Cloudflare introduces Clef and Clef-flash, open-source decision models offering high-speed, structured classification for text and images, available both on Workers AI and to run locally. A new reinforcement learning platform for fine-tuning is also released.
GPT-6.1 Sol offers near-Astra-level intelligence, improved outputs, and a context window over one million tokens—at a fraction of Astra's cost.
Anthropic launches Claude Sonnet 5.5, offering up to 30% faster performance and cost savings compared to Sonnet 5, with a focus on text, interface, and image-related tasks for developers.
ThinkV2V introduces a reasoning-driven framework that leverages multimodal large language models (MLLMs) for complex instruction-guided video editing, outperforming larger baselines and introducing new datasets and benchmarks for evaluation.
MOBA-VL is a new 9B-parameter vision-language model for real-time commentary in MOBA esports. It uses event-localized multi-turn reinforcement learning with supervision from game telemetry, improving accuracy on key moments. MOBA-VL outperforms prior streaming VLMs on new benchmarks, and code, data, and demos are available.
LoopVL is a new vision-language model architecture that leverages recurrent computation through Module-Loop and Model-Loop mechanisms, achieving superior performance and novel interpretability features compared to non-recurrent models.
AREX-2 advances self-improving LLM agents by introducing reflection and long-horizon execution, demonstrating significant gains on both algorithmic and research tasks through iterative solution refinement.
NeurDuo-EEG introduces a new approach to EEG modeling with persistent, multi-timescale memory, significantly improving performance across short and long EEG analysis tasks. The model is open source and supports efficient streaming inference.
The new Function-Space Transformer (FST) architecture introduces spatially adaptive continuous latent representations for function learning, outperforming prior models on PDE prediction and image classification tasks.
Researchers have released DualCast, a dual-path framework combining fast numerical forecasting and optional news-conditioned revisions using a frozen Qwen3-8B language model extended with a custom financial vocabulary. The model achieves state-of-the-art performance across multiple asset classes and forecasting horizons.
