Unverified50% confidenceBenchmarkExact time
Cognition has released coding model Sway 2, achieving 50.0% on the Frontier Code benchmark, claiming to be within 1 point of stronger models while costing 64% less.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/16/2026
First Seen
Valid until: 12/15/2026
Sources
Related Entities
Related Claims
UnverifiedFor 80% of everyday coding tasks, the performance gap between lightweight models and flagship models is within 5%, but the cost difference can be 20–50 times.69% similarUnverifiedHaiku's API call cost is approximately 1/60 that of Opus, while the accuracy gap between the two on simple tasks may be less than 2%.66% similarUnverifiedA tiered model invocation strategy that delegates simple intent recognition to small models in the 7B range and routes complex reasoning to GPT-4o can reduce token costs by more than 60%.64% similarUnverified一个61分的模型和一个59分的模型在多数实际应用中的差异可能远小于分数暗示的程度63% similarUnverified推测解码中当草稿通过率低于50%时，实际加速效果会大打折扣甚至产生负优化63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/927519API
curl https://kongchang.com/api/v1/knowledge/claims/927519MCP
get_claim(id=927519)