Unverified50% confidenceBenchmarkExact time
An anonymous model, Obsidian 3, appeared in Cognition's new Frontier Code 2.0 benchmark, scoring 46.8%, surpassing Opus 5.5's 45.5%.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
11/8/2026
First Seen
Valid until: 11/22/2026
Sources
Gemini 4 Carbon Tops AI Intelligence Leaderboard; DeepSeek V5 and Fable 5.5 Leaked
bilibili一只AI风向标10/29/2026
Related Entities
Related Claims
UnverifiedObsidian 3, a model suspected to be Fable 5.5, scored 46.8% on the Frontier Code 2.0 benchmark at high reasoning intensity, only narrowly leading Opus 5.5's 45.5%.78% similarUnverifiedNuExtract3 scored 0.651 in NuMind's internal test of about 600 documents, versus 0.417 for the base model Qwen3.5-4B; on ExtractBench it averaged 82.35, but neither states coverage of financial report documents.63% similarUnverifiedOn the Frontier Code programming benchmark, the old model scored 34.4%, while Gemini 3.7 Flash improved to 43.6%.61% similarUnverifiedEarly third-party benchmarks show Mistral Large 4 scoring 38 on the Intelligence Index, behind GLM 5.3's 45 and KIMI-K3's 44; its Cyber Index score is 50, and SciBench reaches 93%.60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1147479API
curl https://kongchang.com/api/v1/knowledge/claims/1147479MCP
get_claim(id=1147479)