Bemutatkozott az Inkling 2 Small modell 594 milliárd paraméterrel

Today we are introducing our second model generation with Inkling 2 Small. It is a Mixture-of-Experts model with 594B total parameters, 17B active, and supports image inputs and variable effort. Inkling 2 Small far surpasses our previous models’ scores on the Artificial Analysis Intelligence Index with a score of 31, reflecting gains in performance on agentic coding, vision, knowledge work, and reasoning. The increase in capabilities comes from improvements made at every step of our training process, including more efficient pre-training and a large, unified RL run. Inkling 2 Small will initially be available through Tinker, with closed weights, following the approach we outlined in a safe path to open weights. We intend to make the full weights available under an Apache 2.0 license next week after the completion of external safety testing and red-teaming. You can try Inkling 2 Small on Tinker Playground today. The model is also available for fine-tuning and sampling on Tinker; checkpoints will be downloadable with the release of Inkling 2 Small’s weights. Inkling 2 Small is more capable across the board than our first generation, particularly in long-horizon autonomous work. Inkling 2 Small’s vision and agentic capabilities enable it to incorporate visual feedback to iteratively build and refine websites, games, and 3D designs with interactive elements. Inkling 2 Small was trained with RL on hundreds of thousands of long-horizon coding tasks across multiple agent harnesses and context-compaction strategies. We increased the quality of our synthetic coding data pipeline and moved beyond unit-test rewards alone; many tasks included an agentic RL grader that evaluated how the model approached the problem, not just whether its code passed. We also tested early checkpoints of Inkling 2 Small on OpenCode under the name Fledge-Alpha to assess our model’s performance on coding tasks in the real world. The result is a model that is both stronger at coding and easier to use inside real agent loops. Inkling 2 Small improves on usability compared to Inkling: it follows instructions better, considers changes carefully, and verifies its work without being asked. 69 tasks · Four independent attempts · 10-hour budget Hover a curve or select a model to inspect; select it again to reset. With a model focused, use ← and → to step through results, or Escape to clear. Rows are ordered by AUC: the area under the mean curve over 0–10 hours, with a maximum of 690 task-hours. Each model uses its maximum available reasoning setting, the PI harness and a shared compaction policy. We include fixes for task and grading issues identified during evaluation. Each submitted solution receives two independent LLM audits, with a third review resolving disagreements. Beyond conventional software tasks, Inkling 2 Small was also trained on long-horizon autoresearch tasks. This included agentic GPU kernel optimization and ML tasks such as post-training a model autonomously, all run across a range of GPU architectures. DeepSeek-V3.2 sparse attention, forward + backward To demonstrate its end-to-end autoresearch performance, we tested Inkling 2 Small on a task Proximal built for one of our internal benchmarks. Inkling 2 Small autonomously trained Qwen3.5-4B on Tinker to predict various time series such as weather and Wikipedia page views. It started by analyzing the tasks and writing its own evaluator, experimented with training datasets and recipes, and tried various approaches to SFT and RL, while debugging the process as it goes. After 4.5 hours of work and $426 in Tinker spend, the trained model has more than doubled its forecasting score. Inkling 2 Small processes image inputs using a 292M-parameter vision encoder based on Perception EncoderBolya et al., 2025.. We use 20×20-pixel patches with 2×2 spatial pooling, and interleave image patches with text tokens. The model demonstrates strong visual capabilities across multiple categories, including O