Flint is an open-weight mixture-of-experts model from Recursive, built for coding agents that investigate codebases, fix bugs, and write tests. This page is the Hugging Face model card, covering how to load it and serve it with SGLang and vLLM.

• Total of 39.8B parameters, with roughly 2.5B active per token, and a documented 262,144-token context window.
• Reported scores: 54.9% on Terminal-Bench 2.1, 89.7% on SWE-bench Pro v2, and 74.2% on SWE-bench Verified, with several scores measured by Recursive itself.
• Runs on one B300 or H200 GPU per replica, and the weights are released under Apache 2.0. Transformers modeling code is not included yet.

DeepSeek V4.1 Flash is a 552B-parameter mixture-of-experts model that activates only 8B-16B parameters per token and supports a 1M-token context window for handling long documents and agentic workloads. The model cuts KV-cache memory usage by 4x compared to its predecessor while improving coding, reasoning, and agent performance.

• Activates just 8B parameters during input processing and 16B during generation, making it efficient despite the massive 552B backbone
• Reduces memory footprint by roughly 4x versus DeepSeek-V4-Flash-0731 through Compressed Sparse Attention 2 and a Causal Encoder-Decoder architecture
• Offers configurable reasoning-effort levels so developers can adjust latency and cost versus reasoning depth
• Handles multimodal input (text and images) natively with a 1M-token context window, designed for agents processing large amounts of data

Cohere released North Small Translate, a sparse mixture-of-experts model with 25 billion active parameters designed for machine translation across 50 languages. The model scores 83.60 on WMT26 benchmarks and is available on Hugging Face under a non-commercial license.

• The model uses a sparse MoE architecture with 218B total parameters but only activates 25B per token, reducing computational overhead compared to dense models of similar capability
• It supports 50 languages including major ones like English, Spanish, Mandarin, and Hindi, plus less common ones like Filipino and Maltese
• Multiple quantization options (BF16, FP8, NVFP4) let you run it on different hardware—from 8 H100s down to 2 H100s depending on precision, with production deployment using greedy decoding

The article introduces Inkling-Small, a Mixture-of-Experts transformer with 276 B total and 12 B active parameters that matches full Inkling performance at one-quarter the size. It handles text, images, and audio with a 1 M-token context window and lets users trade compute for thinking effort. The open-weight model outperforms peers on tool use, coding, reasoning, and multimodal benchmarks and is available for fine-tuning and chat.

• Inkling-Small matches or beats full Inkling performance (e.g., 31.6% vs 29.7% on Humanity's Last Exam) while using only 12B active params out of 276B total—roughly a quarter the compute cost.
• It's natively multimodal (text, images via 40×40 patches, audio via dMel spectrograms) with a 1M-token context window and adjustable "thinking effort" for cost/performance tradeoffs.
• Safety numbers slightly lag full Inkling (71.6% vs 78.0% harmful-query refusal on FORTRESS) despite reusing the same safety recipe.
• Open-weight and available now for fine-tuning on Tinker and chat testing in the Tinker Playground.

This article unpacks a 2026 paper arguing that under finite resources, narrowly focused AI systems consistently outperform general-purpose ones. It draws on the no-free-lunch theorem, examples from biology and markets, and machine learning phenomena like negative transfer, mixture-of-experts, and AlphaFold’s task-specific success.

• The no-free-lunch theorem means gains on one task distribution necessarily cost performance on others, so under finite compute/data/time, specialized models win.
• Even massive "general" models rely on mixture-of-experts routing, quietly embedding narrow specialists inside them rather than being truly general.
• Negative transfer in multi-task training shows shared capacity actively creates conflict, dragging down individual task accuracy.
• Biology and markets both confirm the pattern: evolution favors niche specialists over generalists, and focused firms outcompete unfocused ones.

Laguna XS 2.1 is a 33B Mixture-of-Experts model with 3B activated parameters per token, tuned for agentic coding and long-horizon tasks on local machines. It outperforms XS.2 on SWE-bench Multilingual by 5.4 points, adds support for major runtimes, offers FP8/INT4/NVFP4 checkpoints, and runs with a permissive OpenMDW-1.1 license. You can access it via API, OpenRouter, or locally with quantized or DFlash-boosted builds.

• Laguna XS 2.1 jumps from 57.7% to 63.1% on SWE-bench Multilingual, beating Qwen 3.6 (35B) and North Mini Code (30B) despite only 3B active params per token
• DFlash speculator models double token throughput versus stock XS 2.1 when run locally
• Same API pricing as XS.2 ($0.10 input/$0.20 output/$0.05 cache-read per million tokens), now served at 256K context
• XS.2 leaves the API in one week, pushing users to migrate to 2.1 (though it remains available on Baseten's Model Library)