Unverified60% confidenceBenchmarkExact time
In the indirect prompt injection evaluation, the attack success rate against Carbon was 0 (with 15 attempts per attack).
2
Sources
60%
Confidence
Short-term (~14 days)
Relevance
10/29/2026
First Seen
Valid until: 11/12/2026
Sources
Gemini 4 Carbon Tops the Intelligence Leaderboard, But Coding Agents Still Trail Claude Opus 5.5
bilibili进化中的阿陈10/29/2026
Related Entities
Related Claims
UnverifiedIn the experiments, the task resolution rate was 17% for zero-shot generation, 27% for generic self-correction, 23% for error-only feedback, and 90% for diagnostic counterexample feedback (A-CEGIS).69% similarUnverifiedOn two tasks, biomedical inference and vulnerability detection, tested against 18 models, the best-of-10 attack flipped correct answers for an average of 54.9% of cases on proprietary models and reached 80.0% on open-weight models.66% similarUnverifiedINFUSE benchmark results show that the average attack success rate under LLM judging is 19.6% (15.4% under heuristic checks), and prompt-level hardening can reduce the average success rate from 58.2% to 14.3%, but at a utility cost.66% similarUnverifiedIn the AgentDyn benchmark, environmental guidance reduced the attack success rate to 0%.66% similarUnverifiedAccording to Google's safety data, unauthorized shortcuts accounted for 2% of coding evaluation attempts, and indirect prompt injection attacks succeeded in 0.7% of 15 adaptive attempts.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1111467API
curl https://kongchang.com/api/v1/knowledge/claims/1111467MCP
get_claim(id=1111467)