Unverified50% confidenceBenchmarkExact time
In the official side-by-side evaluation, the 1017 version improved by 11.1 points on terminal tasks and 8.8 points on coding ability.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/22/2026
First Seen
Valid until: 1/20/2027
Sources
Related Entities
Related Claims
UnverifiedThe new model improved on all 11 of the vendor's own benchmarks, with Terminal-Bench 4.0 up 11.1 points and ProgramBench up 8.8 points.67% similarUnverifiedOn the code generation task HumanEval, 8-bit scored 92.7%, significantly higher than OQ4E's 89.6%.61% similarUnverifiedAmong the 11 benchmarks officially published, the terminal operation test version 4.0 rose from 31.2 to 42.3, the deep software engineering test version 1.1 rose from 74.2 to 77.0, the automation test reached 60.5, and the Agent ultimate exam scored 35.4.60% similarUnverifiedIn the RULER-16K experiments, KVFetch improved the average score across 13 tasks by +8.4, with gains concentrated on tasks requiring sequential access.59% similarUnverifiedCompared to the strongest baseline, POrchestra improves average task success rate by 9.1%.58% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1086985API
curl https://kongchang.com/api/v1/knowledge/claims/1086985MCP
get_claim(id=1086985)