Harrier-OSS

weights

open(three ungated MIT safetensors checkpoints)

data

closed(over 2B weakly-supervised pairs plus GPT-5 synthetic data, described but not released)

code

partial(inference usage and MTEB evaluation prompts only)

license

MIT(OSI)

An OSI license over ungated weights, with the corpus behind them only narrated. Microsoft describes the pipeline that built the training mixture and the synthetic generation step in prose, and ships neither the data nor the code that produced it, so the family lands on the ladder's open-weights center of gravity rather than a rung above it.

• https://huggingface.co/microsoft/harrier-oss-v1-27b/raw/main/README.md recorded 2026-09-11

  license: mit in the card front matter; three downloadable checkpoints listed; training described as contrastive learning on a large-scale multilingual mixture with distillation, naming no dataset and no training code; usage snippets and an evaluation-prompt file only.

• https://huggingface.co/api/models/microsoft/harrier-oss-v1-27b recorded 2026-09-11

  gated: false; private: false; disabled: false; cardData license mit; sharded safetensors weight files present.

• https://blogs.bing.com/search/April-2026/Microsoft-Open-Sources-Industry-Leading-Embedding-Model recorded 2026-09-11

  Microsoft describes a data pipeline producing more than 2 billion weakly-supervised examples and over 10 million high-quality fine-tuning examples, with GPT-5 used to synthesize multilingual pairs; none of it is offered for download.

765,547 downloads in the trailing 30 days summed across the three shipped checkpoints - 515,530 for the 270m, 243,108 for the 0.6b and 6,909 for the 27b. The split is the shape this category expects: the small on-device sizes carry the volume, and the flagship that tops the benchmark needs serious hardware and is pulled two orders of magnitude less often.

• https://huggingface.co/api/models/microsoft/harrier-oss-v1-270m recorded 2026-09-11

  downloads: 515530; gated: false; private: false.

• https://huggingface.co/api/models/microsoft/harrier-oss-v1-0.6b recorded 2026-09-11

  downloads: 243108; gated: false; private: false.

• https://huggingface.co/api/models/microsoft/harrier-oss-v1-27b recorded 2026-09-11

  downloads: 6909; gated: false; private: false.

The category anchor at rung 5. The 27B checkpoint carries the strongest MMTEB v2 result any publisher has claimed in this category - Microsoft reports 74.3 and first place as of 6 April 2026 - across 100+ languages at a 32k context, and the figure is printed on the model card as well as in the announcement rather than read off a leaderboard screenshot. Every other product in this category is placed against it. The rung rests on the reported 74.3 being the highest figure any publisher claims in this category, not on a verified current leaderboard position: the first-place claim is Microsoft's own, dated 6 April 2026, and its card and its blog are one publisher rather than two independent sources.

• https://huggingface.co/microsoft/harrier-oss-v1-27b/raw/main/README.md recorded 2026-09-11

  Model table gives MTEB v2 scores of 66.5 (270m), 69.0 (0.6b) and 74.3 (27b) at 32,768 max tokens each, and states the models achieve state-of-the-art results on Multilingual MTEB v2 as of the release date.

• https://blogs.bing.com/search/April-2026/Microsoft-Open-Sources-Industry-Leading-Embedding-Model recorded 2026-09-11

  Harrier is described as ranking 1st on the multilingual MTEB-v2 benchmark as of April 6, 2026, supporting more than 100 languages with a 32k context window.