Iris-mini and Iris-pro are the strongest open-weight search agents in their class
Chinese lab AllSpark releases Iris-mini (35B) and Iris-pro (397B) open-weight search agents claiming best-in-class results on BrowseComp and other research benchmarks.
AllSpark's paper introduces Iris-mini (35B parameters, built on Qwen3.6-35B-A3B) and Iris-pro (397B parameters, built on Qwen3.5-397B-A17B), both with 256,000-token context windows. Iris-pro scores 88.6 on BrowseComp, 85.1 on BrowseComp-ZH, 92.9 on DeepSearchQA, and 56.4 on Humanity's Last Exam; Iris-mini reaches 82.2, 84.8, 86.9, and 52.3 respectively. Training tasks are reverse-engineered from web link structure, filtered by a judge model, and refined via alternating SFT and reinforcement learning ('SFT-RL climbing') against live web search. Weights are available on Hugging Face, and the Iris Harness with agent loop, tools, and all four benchmarks is on GitHub.
- Iris-mini (35B) and Iris-pro (397B) built on Qwen-series models with 256k context
- Training questions generated backward from web link graphs so answers require multi-step reasoning
- Iris-pro scores 88.6 on BrowseComp and 56.4 on Humanity's Last Exam
- Context management boosts Iris-mini's BrowseComp score by up to 21.2 points
- Search training unexpectedly improved general tool use and office-task performance
From the original article
4 min read at the-decoder.com
The AllSpark team has released Iris-mini and Iris-pro, two open-source search agents built on Qwen models that lead benchmarks among open-weight models in their size classes. According to the paper, the training data and models also improved performance on tasks they were never trained for, including general tool use and office work.
By Jonathan Kemper · the-decoder.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.