Expired50% confidenceFactExact time
In Microsoft's Copilot Cowork multi-model architecture, Anthropic Opus 4.8 is used for frontier exploration work, Anthropic Sonnet 4.6 for routine processing tasks, GPT 5.5 or the proprietary Cowork 1 for low-cost daily tasks, and DeepSeek on Azure for cost-extremely-sensitive tasks.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/2/2026
First Seen
Valid until: 9/30/2026(expired)
Sources
Related Claims
UnverifiedOn the SWE-bench software engineering task benchmark, the task completion rates of Claude 3.5 Sonnet and GPT-4o far exceed those of open-source models.69% similarUnverified让 Opus 5 扮演架构师制定计划，再把具体执行分配给 Opus、Sonnet、GPT 5.5 或本地模型，可避免在低层次任务上浪费 Token68% similarUnverifiedThe DeepSeek series scores comparably to GPT-4o and Claude Sonnet on some programming tasks in public benchmarks such as HumanEval and SWE-bench.68% similarVerifiedThe top tier of the AI programming assistant market is represented by Claude 3.5 Sonnet and GPT-4o, which excel at complex architectural design and long-context engineering tasks.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/44783API
curl https://kongchang.com/api/v1/knowledge/claims/44783MCP
get_claim(id=44783)