WDCD Run #140 (2026-05-31) evaluated 11 frontier models on multi-turn commitment integrity, finding an average instruction decay of 36.5% from Round 1 to Round 3. Qwen3 Max led the field with 70.8 points and the lowest decay rate at 17%.
In the latest WDCD cycle, all 11 evaluated models show improvement in compliance ability, with the top five all rising and none declining. Gemini 3.1 Pro leaps into the top three with a +14.2 point gain, signaling a major shift in the competitive landscape.
The WDCD compliance test evaluates model stability under real enterprise constraints through three rounds of dialogue. The resource limitation scenario scored the lowest overall, becoming a common "stumbling block" for all 11 models.
Eleven mainstream models showed a clear degradation trajectory in the three-round WDCD test: nearly all confirmed constraints in R1, maintained 93% resistance after R2 interference, but average integrity rate dropped to only 30.5% in R3, with 200 tests directly hitting zero.
The first public ranking of the WDCD compliance test shatters the myth that bigger parameters mean greater reliability. Qwen3 Max leads with 70.83 points, while Grok4 finishes last with 51.67 points; the average crash rate in Phase R3 reaches 60.6%, proving that most models are still highly prone to violating constraints under real enterprise conditions.
Groq, an emerging force in the AI chip field, has recently announced a new funding round and a partnership with Nvidia to jointly expand inference cloud services, reshaping the competitive landscape of AI hardware and cloud infrastructure.
Figure company announced that its third-generation humanoid robot Figure 03 completed a 200-hour continuous operation test, sparking widespread attention in the global robotics field. This achievement injects new momentum into embodied intelligence and drives the integration of AI and robotics into a new stage.
The successful completion of China's Three-Body Computing Constellation marks a new phase in global space AI infrastructure. It achieves full-orbit interconnection with 5P OPS computing power, supporting the operation of 140-billion-parameter large models.
The report presents ten major trends including chip diversification and ultra-large-scale green clusters, highlighting the rise of the Token Economy and strategic implications for nations and enterprises.
According to the "New Generation Artificial Intelligence Technology Industry Development Report 2026", as of the end of 2025, the number of AI enterprises in China exceeded 6,000, and the core industry scale surpassed 1.2 trillion RMB. The year 2026 is defined as a turning point, with large models, agents, and embodied intelligence technologies moving from laboratories to large-scale application.
Anthropic officially launched Claude Opus 4.8 on May 29 and announced the completion of a $65 billion new funding round, reaching a valuation of $965 billion, making it the highest-valued AI company.
This week's Smoke rapid tests over 7 consecutive days reveal DeepSeek V4 Pro's steep decline from 97.08 to 66.88, averaging 79.8 with high volatility. In contrast, GPT-5.5 and Claude Sonnet 4.6 show steady rebounds, with GPT-5.5 rising 11.5 points.
Meta has been revealed to deploy a mouse tracking tool internally to monitor employee behavior, sparking heated debate over privacy and efficiency. The tool, aimed at optimizing remote work, raises concerns under the EU's General Data Protection Regulation (GDPR).
A discussion about Claude's simulated portfolio has sparked industry debate, as it buys ServiceNow, viewing the company as a beneficiary of AI agent infrastructure rather than a victim, leading to a noticeable stock rebound.
Oppo has announced the open-sourcing of its X-OmniClaw Android AI agent framework, a significant breakthrough in on-device AI. The framework emphasizes local data processing to avoid privacy risks from cloud transmission while supporting multimodal perception and autonomous decision-making.
Senator Elizabeth Warren's AI tax proposal has sparked intense debate among tech and political circles, with an estimated annual revenue of $4 trillion aimed at funding social programs.
NVIDIA and Dell recently showcased the AI Factory solution at TechWorld, drawing widespread industry attention. The solution aims to help enterprises deploy on-premises agentic AI systems while supporting the integration of physical robots, marking an extension of AI technology from the cloud to edge and local environments.
Google has rolled out a major update in AI search, advancing its Agentic AI Search strategy by introducing intelligent information agents and multimodal processing capabilities, showcasing the latest progress of the Gemini model series. This move is seen as a critical step in transforming search technology from passive response to active agency.
Microsoft is accelerating the transformation of Copilot into a super app, integrating scattered AI tools into a unified workspace. This move could reshape enterprise automation while also facing regulatory scrutiny and competition.
Anthropic has announced a major update to Claude Opus 4.8, focusing on enterprise applications by introducing dynamic system instruction injection within conversations, significantly optimizing caching mechanisms and API call performance.
