Unverified50% confidenceBenchmarkExact time
About half of the tasks are easy for all models, with scores close to 100%; on the other half of difficult tasks, the three models scored 50%, 45%, and 23% respectively.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
10/6/2026
First Seen
Valid until: 10/20/2026
Sources
Related Entities
Related Claims
UnverifiedA WANDR score of 0.610 does not mean the model fully succeeds on 60% of tasks; the actual partial completion rate is related to the distribution of task complexity.69% similarUnverifiedA sensible model selection strategy is to use Flash to handle 80% of simple tasks and only invoke the flagship model for the remaining 20% of complex scenarios.68% similarUnverifiedApproximately 60–70% of actual production requests in enterprises are simple tasks, and for such tasks, models with fewer than 7B parameters can achieve output quality highly similar to flagship models, while the cost difference is as much as 20–50 times.68% similarUnverifiedProviding 2 to 5 typical examples (few-shot) can significantly improve model performance on complex tasks compared to zero-shot.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/980437API
curl https://kongchang.com/api/v1/knowledge/claims/980437MCP
get_claim(id=980437)