88Z AI, 26 May — Kunlun Wanwei's Tiangong AI releases SkyClaw-v1.0 and its lite version, supporting million-token context, optimized for complex tool use and code generation, outperforming mainstream open-source models at a lower price.
At a glance
- SkyClaw-v1.0 outperforms mainstream open-source models like Minimax 2.7 in agent benchmarks, approaching top-tier model performance.
- The model supports million-token context, optimizes complex tool use and code generation, and is priced at less than half of competitors.
The update
Kunlun Wanwei Group's Tiangong AI has officially launched the high-performance agent model SkyClaw-v1.0 and its lightweight version, SkyClaw-v1.0-lite. These models support long contexts of up to one million tokens and are deeply adapted for real-world agent workflows. They feature targeted optimizations for complex tool invocation, multi-turn task execution, code generation, and interactive application building. In multiple agent benchmarks, SkyClaw-v1.0 comprehensively outperforms mainstream open-source models such as Minimax 2.7 and DeepSeek V4 Flash. Its performance approaches that of larger top-tier models, while its pricing is highly competitive, set at less than half the cost of competitors, aiming to provide a cost-effective solution for agent development.
Sources & editorial note
This article is an AI-assisted synthesis by 88Z AI based on the public sources below. Product claims and availability are attributed to their publishers. It does not represent independent testing or first-hand reporting.
- 昆仑万维集团Source published
Source dates above refer to the original announcements; the date under the headline is this article’s publication on 88Z AI.