Inception42 reports that the post-graph-rag model achieved 98.4% temporal reasoning performance compared to 62.4% on Zep/Graphiti and 60.2% on full-context GPT-4o.
Public source
Publisher name
Public post
𝗧𝗵𝗲 𝗽𝗼𝘀𝘁-𝗴𝗿𝗮𝗽𝗵-𝗿𝗮𝗴 𝗽𝗮𝗽𝗲𝗿 𝗶𝘀 𝗻𝗼𝘄 𝗼𝗻 𝗮𝗿𝗫𝗶𝘃 — 𝗮𝗻𝗱 𝗼𝗻 𝗟𝗼𝗻𝗴𝗠𝗲𝗺𝗘𝘃𝗮𝗹, 𝘁𝗵𝗲 𝗯𝗲𝗻𝗰𝗵𝗺𝗮𝗿𝗸 𝗭𝗲𝗽 𝗽𝘂𝗯𝗹𝗶𝘀𝗵 𝗳𝗼𝗿 𝗚�…
Company
Inception42
- Industry
- Technology, Information and Internet
- Location
- Abu Dhabi, AE
- Company size
- 201–500 employees
About Inception42
Our mission is to accelerate the ambitions of organisations and nations by turning intelligence into impact. AI native. Human led.
See moreLatest activity
Latest activity from Inception42
7 signals
Presence & Recognition
Inception42 CEO Ibrahim AlJallaf and senior technology leaders across industries will convene in a closed-door discussion on Unlocking AI for the GCC Workforce at the Innovation Hub, Festival Tower, and Dubai Festival City on September 8, 2026.
Presence & Recognition
Inception42 CEO Ashish Ipe Koshy will join global leaders at AIM Congress for the Digital Economy Leaders' Panel exploring strategic investments needed to build competitive and resilient AI ecosystems.
Products & Services
Inception42 launched Catalyst to help organizations turn Agentic AI aspirations into operational impact.
Discover more
Similar signals
Similar public activity from other companies.
Research & Knowledge
Turing
Turing announced that GPT-5.6 Terra scored highest of the six models tested, averaging just 41% of the expert-written rubric criteria across the 40 tasks.
Research & Knowledge
SinAura
SinAura rethinks how AriaPLT reasons before acting by building on a depth-recurrent transformer architecture that loops through a shared recurrent block multiple times to deepen internal reasoning before producing a token.
Research & Knowledge
NEC Laboratories America, Inc.
NEC Laboratories America, Inc. published a paper titled Stop When Enough: Adaptive Early-Stopping for Chain-of-Thought Reasoning on July 8, 2026.
Research & Knowledge
Towards Data Science
Towards Data Science published a research paper comparing cost, latency, and answer quality using a million-token context window and Kimi K3 with a RAG pipeline.
Research & Knowledge
LlamaIndex