{"version":1,"type":"story","url":"https://digestai.news/story/tinyleague-updates-benchmarks-drops-aa-index-adds-kv-efficiency-metric","json":"https://digestai.news/story/tinyleague-updates-benchmarks-drops-aa-index-adds-kv-efficiency-metric.json","markdown":"https://digestai.news/story/tinyleague-updates-benchmarks-drops-aa-index-adds-kv-efficiency-metric.md","slug":"tinyleague-updates-benchmarks-drops-aa-index-adds-kv-efficiency-metric","headline":"TinyLeague updates benchmarks: drops AA Index, adds KV efficiency metrics","summary":"TinyLeague, a tracker for small open-weight AI models, updated its benchmarking platform on October 28, 2026. The latest changes include removing the AA Intelligence Index and detail tables, while adding **KV cache efficiency** metrics and stricter cutoff criteria. The platform now focuses on **KV cache requirements** (measured at 128K tokens) and rates models on a **0–100 log scale** for efficiency, excluding hybrid-layer overheads. Two models were unrated due to missing active-parameter data, and the heatmap now hides columns with fewer than two entries when filters are applied.\n\nKey additions include **LiquidAI/LFM3-1.7B**, **Sage-A1.5**, and **XHToken/Spark-X2.6-4B**, alongside vision-capable models like **Ornith-1.5-122B** and **MiniCPM-V-4.7-35B**, the latter noted as the first open-source on-device MoE VLM. Older entries like **Nemotron-3-Nano-Omni-30B-A3B-Reasoning** were removed. The updates aim to improve clarity and comparability for researchers evaluating small models.","keyPoints":["Removed AA Intelligence Index and detail tables; added KV cache efficiency metrics (0–100 scale)","Two models unrated due to missing active-parameter data; heatmap hides low-entry columns under filters","Added 10+ new models, including LiquidAI/LFM3-1.7B and MiniCPM-V-4.7-35B (first open-source on-device MoE VLM)"],"whyItMatters":"KV cache efficiency is critical for deploying small models at scale, balancing cost and performance. This update helps researchers pick models that minimize memory overhead, accelerating adoption of lightweight yet capable AI.","category":{"slug":"research","name":"Research","url":"https://digestai.news/category/research"},"entities":{"companies":["TinyLeague","LiquidAI","NVIDIA","openbmb","IFM","Agnes-AI"],"models":["LFM3-1.7B","Sage-A1.5","Spark-X2.6-4B","Ornith-1.5-122B","LFM2.5-24B-A2B","NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16"],"people":[]},"firstPublishedAt":"2026-10-28T12:59:19Z","updatedAt":"2026-10-28T12:59:19Z","sourceCount":1,"hasPrimarySource":false,"sources":[{"outlet":"tinyleague.nunodonato.com","title":"Important update to TinyLeague (AA out, KV efficiency in)","url":"https://tinyleague.nunodonato.com/","publishedAt":"2026-10-28T12:59:19Z","type":"press","primary":false,"lead":true}],"sourceNotes":null,"discussions":[{"site":"Reddit","url":"https://www.reddit.com/r/LocalLLaMA/comments/1xhndti/important_update_to_tinyleague_aa_out_kv/","points":null}],"thread":null,"cite":{"text":"Digest AI, \"TinyLeague updates benchmarks: drops AA Index, adds KV efficiency metrics\", 28 October 2026, https://digestai.news/story/tinyleague-updates-benchmarks-drops-aa-index-adds-kv-efficiency-metric","publisher":"Digest AI","title":"TinyLeague updates benchmarks: drops AA Index, adds KV efficiency metrics","datePublished":"2026-10-28T12:59:19Z","url":"https://digestai.news/story/tinyleague-updates-benchmarks-drops-aa-index-adds-kv-efficiency-metric"},"generatedBy":"Written by Digest AI's editorial model from the linked sources; the sources are the record.","license":"Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse"}