Ornith-1.5: From Self-Scaffolding to Self-Improvement | Ornith Blog

Today, we are introducing Ornith-1.5, a major step toward building foundation models through end-to-end self-improvement. Ornith-1.5 extends the self-scaffolding framework<br>introduced in Ornith-1.0 into a more complete self-improvement loop: the model<br>proposes new tasks, generates task-specific scaffolds, and produces solution rollouts for reinforcement learning, continuously creating new learning<br>experiences from which it can improve.

Ornith-1.5 spans three model scales: 397B MoE, 35B MoE and 9B dense. Designed for strong general-purpose intelligence across reasoning, agentic, and coding tasks, Ornith-1.5<br>achieves state-of-the-art performance among open-source models of comparable size across a broad range of benchmarks. Ornith-1.5-397B scores 86.1 on Terminal-Bench 2.1 and 56.0<br>on DeepSWE, performing on par with Claude Opus 4.8 (85.0 and 59.0) while outperforming leading open-source models of similar scale, including GLM-5.2 (82.7 and 46.2) and<br>DeepSeek-V4-Flash-0731 (82.7 and 54.4). At the other end of the spectrum, Ornith-1.5-9B, with its quantized Ornith-1.5-9B-Mobile version, can be readily deployed on iPhone and<br>Android devices while substantially outperforming larger models such as Gemma 4-31B and Qwen 3.6-35B.

At the flagship scale, Ornith-1.5-397B achieves 86.1 on Terminal-Bench 2.1 and 56 on DeepSWE, matching Claude Opus 4.8 on both benchmarks and outperforming leading<br>open-source models of similar size, including GLM-5.2 and DeepSeek-V4-Flash-0731.

Ornith-1.5-35B significantly outperforms its similar-sized peer Qwen 3.6-35B across all coding and agentic benchmarks, and despite activating only 3B parameters per token,<br>it also outperforms dense models—Gemma 4-31B and Meta’s Muse Glimmer-30B—by wide margins on agentic coding (68.5 vs. 43.4 and 51.7 on Terminal-Bench 2.1; 79.0 vs. 52.0 and<br>76.0 on SWE-Bench Verified).

The edge-deployable Ornith-1.5-9B also delivers remarkably strong results, achieving 47.0 on Terminal-Bench 2.1 and 70.6 on SWE-Bench Verified. Despite being a compact<br>9B-parameter model, it matches or exceeds the performance of much larger models such as Gemma 4-31B and Qwen 3.6-35B.

Self-Improvement through Self-Generated Tasks, Harnesses, and Solutions

Ornith-1.5 extends Ornith-1.0 by expanding the self-improvement loop from scaffold and rollout optimization to jointly optimizing task generation, scaffold construction, and<br>solution rollouts. Rather than relying on a fixed set of human-curated tasks and manually designed harnesses, Ornith-1.5 continuously generates new training tasks, discovers<br>effective strategies for solving them, and improves the policy through reinforcement learning.

Each training cycle proceeds in three stages. Given an environment or codebase, high-level instructions about the task type, and access to the model’s previous task-solving<br>history, the system proposes progressively harder tasks that go beyond what the model has already solved, exposing capability gaps and continuously pushing the training<br>frontier.

For each task, the model then generates or refines a task-specific scaffold—the instructions, tools, decomposition strategy, and orchestration used to approach the problem.<br>Conditioned on the task and scaffold, the policy produces a solution rollout. Reward from the rollout is propagated across all three stages, so the system learns not only to<br>produce better solutions, but also to generate more useful training tasks and construct more effective scaffolds.

Repeated over training, this creates a closed self-improvement loop in which stronger policies enable the generation of harder and more informative tasks, evolving scaffolds<br>discover better ways to elicit the model’s capabilities, and higher-quality rollouts provide increasingly effective learning signals. Instead of relying on a static training<br>distribution or hand-engineered agent design, Ornith-1.5 continually expands its own curriculum and adapts its problem-solving strategies, driving sustained capability gains<br>across reasoning, coding, and agentic tasks .

Task Reward

For the question → scaffold → rollout setup, we define the task reward using three signals: validity, frontier difficulty, and novelty . Let<br>\(q\) denote a generated question, \(s\) its scaffold, and \(\{\tau_i\}_{i=1}^{N}\) a set of solution rollouts. We define

\[ R_{\text{task}} = \underbrace{V(q,s)}_{\text{Is it valid and verifiable?}} \times \underbrace{D\!\left(q,s,\{\tau_i\}_{i=1}^{N}\right)}_{\text{Is it at the right<br>difficulty?}} \times \underbrace{N(q)}_{\text{Is it sufficiently novel?}}. \]

Here, \(V\) measures whether the generated task and scaffold form a valid and verifiable learning environment, \(D\) measures whether the task lies near the model’s current<br>capability frontier based on rollout performance, and \(N\) measures novelty relative to previously generated or trained-on tasks. The multiplicative formulation encourages the<br>proposer...