Evidence100%Authoritative
FactConfirmedProduct·March 20, 2026
GPT-5.3-Codex-Spark Ships as First Model Running on Cerebras WSE-3 for OpenAI
OpenAI released GPT-5.3-Codex-Spark, described as the first OpenAI model to run on Cerebras' WSE-3 chip, delivering 25% faster output and ~50% token reduction compared to earlier models, marking the first milestone of the OpenAI-Cerebras partnership.
Evidence Strength
Evidence100%Authoritative
Backed by official company doc
Reported by 2 independent publishers
Includes official or primary source
Key Development
Transformative development (significance 9-10)
Confirmed — verified event
Insights
First tracked
February 12, 2026
Last updated
March 20, 2026
Sources
2 sources
Related Developments
Cerebras Delivers 3,000 Tokens/Second Inference for OpenAI's gpt-oss-120B Open-Weight ModelOpenAI Signs $10B+ Multiyear Compute Deal with CerebrasOpenAI GPT-5.3-Codex-Spark Powered by Cerebras Launches in Research PreviewOklahoma City AI Datacenter Ribbon-Cutting with 44+ ExaflopsCS-3 vs. NVIDIA DGX B200 Blackwell Benchmarks Published
Sources (2)
Source Timeline
Why the AI Race Shifted to SpeedCerebras·Mar 20
OpenAI’s rapid GPT-5.3-Codex model moves beyond simple coding tasksSiliconANGLE·Feb 12
Evidence Strength
Evidence100%Authoritative
Backed by official company doc
Reported by 2 independent publishers
Includes official or primary source
Key Development
Transformative development (significance 9-10)
Confirmed — verified event
Insights
First tracked
February 12, 2026
Last updated
March 20, 2026
Sources
2 sources
Related Developments
Cerebras Delivers 3,000 Tokens/Second Inference for OpenAI's gpt-oss-120B Open-Weight ModelOpenAI Signs $10B+ Multiyear Compute Deal with CerebrasOpenAI GPT-5.3-Codex-Spark Powered by Cerebras Launches in Research PreviewOklahoma City AI Datacenter Ribbon-Cutting with 44+ ExaflopsCS-3 vs. NVIDIA DGX B200 Blackwell Benchmarks Published