CMSManhattan·released 2026-08-031 source
Not yet benchmarked
Claims drawn from cited facts, not live model generation.
Developed by CMSManhattan and released in 2026, this model is a mid-size system with roughly 14.77 billion parameters — a scale that trades some of the efficiency and low cost of smaller models for meaningfully more capability headroom, at a correspondingly higher serving cost.
3 cited facts
This model is not yet benchmarked, as it is currently scored on none of the tracked benchmarks. Because it appears on no tracked benchmarks at all, it also holds no current top score.
2 cited facts
Because the average gap to the state of the art is not disclosed, the only positional reading available is leaderboard standing rather than a numeric distance from the best score. This model holds the top score on none of its benchmarks, which means there is no genuine category leadership to point to yet.
2 cited facts
2 facts cross-checked across data sources: 2 single-source.
Source facts, citations, and refresh stamp for this record.
Sources: huggingface_models
JiRackUltra_14b is an AI model developed by CMSManhattan, released 2026-08-03. Its benchmark record below tags every score as independently reproduced, self-reported, or unverified.