Unverified50% confidenceBenchmarkExact time
In OpenAI's Agent's Last Exam benchmark, Terra at the medium reasoning level scored 15 percentage points higher than the previous generation at one-third the cost, performed close to Astra, and was priced at one-tenth of Astra per token.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
11/1/2026
First Seen
Valid until: 1/30/2027
Sources
Related Entities
Related Claims
UnverifiedOpenAI claimed via The Next Web that Terra completes tasks at about one-ninth the cost of Claude Opus 5.5 and makes fewer factual errors than the model it replaces, but these are all OpenAI's own test results and lack third-party reproduction.64% similarUnverifiedArtificial Analysis gave GPT-6.1 Sol a score of 52, only one point below Astra, with benchmark run costs at one-quarter of Astra's and 31% cheaper than 6 Sol.64% similarUnverifiedOn LiveBench long-horizon agentic tasks, Opus and Astra both score above 80, while Argon scores only about 70.62% similarUnverifiedIn a test checking whether AI-modified code could be accepted by a project, Opus 5.5 using the default setting scored slightly higher than Astra's best publicly available score, and the cost per task was approximately one-fifth of Astra's.61% similarUnverifiedGPT-6 Astra achieved a near-perfect score of 99.9% on ARC-AGI-type tests, while the average score of ordinary human test-takers was only 48%.57% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1121978API
curl https://kongchang.com/api/v1/knowledge/claims/1121978MCP
get_claim(id=1121978)