Ornith-1.0-35B

Ornith AI released Ornith-1.0-35B, a 35B-parameter mixture-of-experts model with 3B active parameters per token, on June 25, 2026, under an MIT license on HuggingFace, featuring a 262K context window. The model reports coding scores of 64.2 on Terminal-Bench 2.1, 75.6 on SWE-bench Verified, and 50.4 on SWE-bench Pro, reasoning scores of 20.8 on HLE and 86.2 on GPQA-Diamond, and agentic scores of 64.4 on MCP-Atlas, 42.4 on Toolathlon-Verified, and 69.8 on ClawEval. It has been superseded by Ornith-1.5-35B-A3B, which achieves higher scores (TB-2.1 67.8, SWE-bench Verified 79.0), and is available via GGUF quantizations for local use on enthusiast-class GPUs.

Ornith-1.0-35B MoE enthusiast ~35B total, ~3B active per token MoE - the Ornith-1.0 enthusiast size, released 2026-06-25 and superseded by Ornith-1.5-35B-A3B. 262K context, MIT-licensed on HuggingFace at ornith-ai/Ornith-1.0-35B GGUF and FP8 quantizations . - Coding vendor self-reported : Terminal-Bench 2.1 64.2, SWE-bench Verified 75.6, SWE-bench Pro 50.4, NL2Repo 34.6. - Reasoning: HLE 20.8, GPQA-Diamond 86.2. - Agentic: MCP-Atlas 64.4, Toolathlon-Verified 42.4, ClawEval 69.8. Superseded by Ornith-1.5-35B-A3B TB-2.1 67.8, SWE-bench Verified 79.0 . Prefer 1.5 for new work. Runs locally via GGUF on enthusiast-class GPUs. coding reasoning agentic - 35.0B - 262k - mit - 🇺🇸 USA - Jun 2026 Scores Coding 86 Reasoning 84 Tool calling 80 General 82 Or run it in the cloud No per-token API provider pricing tracked for Ornith-1.0-35B yet. For flagship list prices, see the calculator /calculator . PRICE HISTORY Inference cost over time Data accumulates from the first daily sync - longer ranges populate over time. Prices come from OpenRouter snapshots, not a historical API. Loading price history...