open-weight model

Bilingual

BTL-3: A 27B open-weight agent model for agentic coding and structural tool use
2 months ago
- BTL-3 is a 27B open-weight agent model designed for agentic coding, structural tool use, repository work, failure recovery, and long multi-turn execution.
- It achieves high scores on benchmarks: 95.1% HumanEval pass@1, 88.5% BFCL v4 AST, and 88.1% LiveCodeBench v6.
- A Compact edition (8.39 GB) packages the full model into a single file with 92.2% behavior retention, enabling native local inference.
- The model is a PEFT LoRA adapter (rank 32, alpha 64) based on Qwen3.6-27B, with a 262,144-token context window and Apache-2.0 license.
- It can be used with Transformers (via PEFT) or vLLM, with recommended non-thinking mode for stability.
- Intended use cases include coding, debugging, test-driven repair, structured function calling, and long multi-turn tasks requiring verification and recovery.
- Operational guidance recommends sandboxing generated code and requiring confirmation for high-impact actions.
Solar Open 2: Korea's Sovereign Foundation Model, Built for Agentic Use
2 months ago
- Solar Open 2 is an open-weight foundation model optimized for agentic use, designed to serve as a base model for agents in real work environments.
- It uses a Mixture-of-Experts (MoE) architecture with 250B total parameters but activates only 15B per token, improving both capability and inference efficiency.
- The model supports a 1M-token context window and runs on two NVIDIA H200 GPUs with quantization, making it practical for enterprise deployment.
- It officially supports Korean, English, and Japanese, with a tokenizer optimized for Korean efficiency (50–80% of tokens compared to global models).
- Training data focused on agent scenarios including search, tool calling (MCP), coding, and office work, with a custom data synthesis and verification pipeline.
- Solar Open 2 achieved top scores among comparable models on MMLU-Pro (86.2), LiveCodeBench (92.4), and Korean benchmarks (85.4 average).
- On Korean office work benchmark Ko-GDPval, it scored 86.8, nearly matching a model six times larger (DeepSeek-V4-Pro) while using only 15B active parameters.
- The model uses Selective Weight Transfer from Solar Open 100B for efficient training, reaching target loss with 58% of the tokens needed by random initialization.
- Weights are released on Hugging Face under a commercially usable license, with a technical report covering architecture, training methodology, and evaluation conditions.
- Solar Open 2 is built for long-horizon tasks, long-context processing, and reliable tool calling, addressing the unique demands of agentic workflows.
Gemma 4 on Cerebras - The Fastest Inference Is Now Multimodal
3 months ago
- Gemma 4 31B runs at over 1,800 tokens per second on Cerebras Inference, enabling fast multimodal applications.
- It is the first Google DeepMind model on Cerebras with image input support for screenshots, documents, charts, and UI states.
- Cerebras achieves record speed and low latency, making Gemma 4 suitable for real-time visual and agentic workflows.
- Gemma 4 31B matches Claude Haiku 4.5 in intelligence but runs 18x faster and is open-weight under Apache 2.0.
- The model's high speed transforms product development, allowing instant iteration and complex multimodal loops.
- As a dense, efficient model, Gemma 4 31B is designed for quality without the large footprint of MoE models.
- Example applications include real-time screenshot analysis, long-context summarization, and UI debugging.
- Gemma 4 31B is available on the Cerebras Inference Cloud in public preview for multimodal workloads.
Kimi K2.7 Code is generally available in GitHub Copilot
3 months ago
- Kimi K2.7 Code, an open-weight model, is now generally available in GitHub Copilot as a selectable option, offering more choice and lower cost.
- It is hosted on Microsoft Azure and billed at provider list pricing under usage-based billing.
- Rolling out initially to Copilot Pro, Pro+, and Max plans, with expansion to Business, Enterprise, and other surfaces planned over coming weeks.
- Available in various tools including Visual Studio Code, Visual Studio, Copilot CLI, GitHub Copilot app, and mobile platforms.
- For Copilot Business and Enterprise, administrators must enable the Kimi K2.7 Code policy in settings, as it is off by default, with recommendations to review security and compliance requirements.

open-weight model

BTL-3: A 27B open-weight agent model for agentic coding and structural tool use

Solar Open 2: Korea's Sovereign Foundation Model, Built for Agentic Use

Gemma 4 on Cerebras - The Fastest Inference Is Now Multimodal

Kimi K2.7 Code is generally available in GitHub Copilot