Is Ornith-1.5-122B state of the art?Pending benchmarks.Announced Oct 20, 2026; no independent results yet.
Ornith's 122B mixture-of-experts model for coding and agentic work, with open weights under the MIT license. The lab says it runs on a single NVIDIA DGX; the model card says a single 80GB GPU.
The details so far
- Lab
- Ornith
- Announced
- Oct 20, 2026
- API
- Not yet
- Open weights
- Published
- Parameters
- 122B MoE (lab); ~125.1B in Hugging Face safetensors (BF16)
- Lab
- Ornith (Hugging Face org ornith-ai)
- Architecture
- Mixture of experts; Hugging Face config: qwen3_5_moe, 256 experts, 8 per token
- Context
- 262,144 max position embeddings (config)
- Weights license
- MIT (Hugging Face card)
- Other weights
- FP8, GGUF and MLX versions on Hugging Face
- API
- Not mentioned
Tags marked unsure are not stated on the lab's own pages.
What would let us call it
Verdicts use only independent boards we already track. Two consecutive snapshots are required before any crown.
- Publicly announced with a source
- Usable by the public (API or weights)
- Listed on an independent board we track
- Enough results for a rank in a category
- Holds across two snapshots (crowns only)
What the lab claims
Self-reported. Shown for context; not used in any verdict.
Lab claim (announcement post): SWE-bench Verified 0.845, Terminal-Bench 2.1 0.813, HLE 0.446, BrowseComp 0.785. Self-reported; not results on isitsota boards.
The Hugging Face card's own table lists Terminal-Bench 2.1 at 81.2 and HLE at 48 (with tools) / 37.68 (without tools), which differ from the post.
Lab claim: state-of-the-art performance among open-source models of comparable size.
Sources
LABOrnith announcement on Xx.com/ornith_/status/2112553477457268799 · Oct 20, 2026REPOOrnith-1.5-122B on Hugging Facehuggingface.co/ornith-ai/Ornith-1.5-122B · Oct 16, 2026Tracking continues until an independent board lists it.