Unverified50% confidenceBenchmarkExact time
On the bug-fixing question, Terra scored full marks in all three tests and fixed the hidden bug at its root, whereas Opus and Argon fixed only half of it on that question.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
11/1/2026
First Seen
Valid until: 11/15/2026
Sources
GPT 6.1 Terra Hands-On: Three Times Faster, Half the Cost, but Prone to These Mistakes
bilibili我是张大宁10/27/2026
Related Entities
Related Claims
UnverifiedIn project three (Lakebed product audit), Gemini 4 Argon scored only 2.25/10, with 56 of 78 findings being wrong, a failure rate of about 70%; while 6.1 Sol scored 9/10.60% similarUnverifiedIn 4 agentic coding tests, Argon won two, and ranked last in both the Frontier SWE and Terminal Bench tests.58% similarUnverifiedArgon tied for first place with the highest score of 68% on CWE-bench v1, which evaluates vulnerability remediation capability.54% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1122009API
curl https://kongchang.com/api/v1/knowledge/claims/1122009MCP
get_claim(id=1122009)