Ant Group's Ling-2.6-flash Model Open-Sourced: 104B Parameters With 7.4B Active, Achieves Multiple SOTA Benchmarks

Gate News message, April 29 — Ant Group’s Ling-2.6-flash model weights are now open-sourced, having previously been available only via API. The model features 104 billion total parameters with 7.4 billion activated per inference, a 256K context window, and MIT licensing. BF16, FP8, and INT4 precision versions are available on HuggingFace and ModelScope.

Ling-2.6-flash introduces hybrid linear attention improvements over Ling 2.0, upgrading the original GQA to a 1:7 MLA plus Lightning Linear hybrid architecture combined with highly sparse MoE. Inference efficiency significantly exceeds comparable models: peak generation speed reaches 340 tokens/s on 4x H20 GPUs, with prefill and decode throughput approximately 4x higher than comparable open-source models. Agent-related benchmarks show strong performance: BFCL-V4, TAU2-bench, SWE-bench Verified (61.2%), Claw-Eval, and PinchBench achieve or approach SOTA levels. Across the full Artificial Analysis benchmark suite, total token consumption is only 15 million. On AIME 2026, the model scored 73.85%.

Ant Group’s official website also lists Ling-2.6-1T (trillion-parameter flagship version) and Ling-2.6-mini (lightweight version), though as of publication, their weights remain unreleased on HuggingFace, with only the flash series available for download.

Disclaimer: The information on this page may come from third parties and does not represent the views or opinions of Gate. The content displayed on this page is for reference only and does not constitute any financial, investment, or legal advice. Gate does not guarantee the accuracy or completeness of the information and shall not be liable for any losses arising from the use of this information. Virtual asset investments carry high risks and are subject to significant price volatility. You may lose all of your invested principal. Please fully understand the relevant risks and make prudent decisions based on your own financial situation and risk tolerance. For details, please refer to Disclaimer.

Related Articles

AI Financial Platform Rogo Raises $160M in Series D Led by Kleiner Perkins in Less Than 3 Months

According to Beating, AI platform Rogo designed for high-frequency financial scenarios completed a $160 million Series D funding round in April 2026, led by Kleiner Perkins with participation from Sequoia, Thrive Capital, Khosla Ventures, and J.P. Morgan. The funding came less than three months

GateNews4h ago

China Blocks Meta-Backed Manus AI Acquisition on April 29, Citing Tech and Data Security Concerns

According to PANews, on April 29, China's National Development and Reform Commission investment security review office banned a foreign acquisition of the Manus project and required the transaction be terminated. Manus, billed as the world's first general artificial intelligence agent, had

GateNews4h ago

Alibaba Cloud Cuts DeepSeek-V4-Pro Implicit Cache Pricing to 1 Yuan per Million Tokens on April 29

According to Alibaba Cloud, its Bailian platform will reduce the pricing for DeepSeek-V4-Pro model's implicit cache (Implicit Cache) to 1 yuan per million tokens effective April 29, 2026 at 23:59:59 Beijing time. Implicit cache only applies when requests hit the cache; cached input tokens are

GateNews5h ago

AI Platform Certifyde Raises $2M in Seed Funding with Ripple CEO Brad Garlinghouse

According to ChainCatcher, AI application platform Certifyde announced the completion of a $2 million seed funding round. Investors include K5 Global, Flamingo Capital, and angel investors such as Ripple CEO Brad Garlinghouse, Honey co-founder George Ruan, and Nutra co-founder Roland

GateNews7h ago

DeepSeek Launches Image Recognition Feature in Beta Testing

According to PANews, DeepSeek launched its image recognition feature today (April 29), currently in beta testing. Both the web version and mobile app users may be selected for the beta rollout.

GateNews8h ago

Anthropic Launches 8 Creative Tool Connectors for Claude, Including Blender, Adobe, Autodesk

Anthropic has announced a suite of creative tool connectors that enable Claude to directly control professional software used by designers and musicians. The initial eight connectors span 3D modeling, visual design, music production, and live performance, with partners including Blender, Adobe,

GateNews8h ago
Comment
0/400
No comments