Arctic AWM 4B

AstaBrief

Allen Institute for AI

AstaBrief is Ai2’s 8B scientific report-generation model, trained from Qwen3-8B with citation-focused supervised fine-tuning and direct preference optimization. It turns a research question and retrieved literature excerpts into a complete cited report in a single pass, and powers Asta’s Fast report mode. Ai2 releases the weights, training data, and a workflow for using local PDFs. The reported latency advantage evaluates the complete report pipeline rather than the checkpoint alone. Citation grounding is improved but does not guarantee that a synthesis preserves every study’s scope or conclusions.

MiniCPM5

OpenBMB

MiniCPM5 is OpenBMB’s fifth-generation dense language-model family for local and resource-constrained use. The 1.08B model was announced May 19, followed by a 2.52B model on September 7. Both support a documented 128K context and target reasoning, coding, tool use, writing, and instruction following. The training recipe combines high-quality data, instruction tuning, and complementary reinforcement-learning signals. Apache-licensed base, intermediate, supervised, and final checkpoints are grouped as variants, alongside routine deployment formats. Small parameter counts improve local practicality but do not eliminate errors or the need for application-specific evaluation.

SERA

Allen Institute for AI

SERA is Ai2’s first Open Coding Agents model family, trained to make repository-level software changes. Soft Verified Generation constructs supervised agent trajectories without requiring every trace to pass strict execution verification, reducing training cost while preserving useful development behavior. Released 8B, 14B, and 32B checkpoints include general and repository-adapted variants. The main 32B checkpoint derives from Qwen3-32B and uses a 32K task context. SWE-bench outcomes depend on the coding harness and environment. The research models do not include general safety filtering and require oversight for executing generated code.