As of 2026-08-12, Cerebras Systems Inc. (CBRS) disclosed AI-agent language in 3 SEC filings since 2026-05-04 (source: Agentic 10-K, from SEC EDGAR).
“• Cerebras fast inference serves as the foundation for agentic flows in industries ranging from finance to life sciences with customers including Block, Figma, AlphaSense, and GSK.”
“Reasoning Drives the Inference Compute Explosion Smarter Models Require More Compute During Inference Tokens = Compute Single Shot Inference PT -4o Reasoning Models GPT-5 Agentic + Reasoning Codex Agentic models orchestrate multiple models, each performing search, reasoning, and synthesis.”
None identified
None named
“• Cerebras fast inference serves as the foundation for agentic flows in industries ranging from finance to life sciences with customers including Block, Figma, AlphaSense, and GSK.”
“Reasoning Drives the Inference Compute Explosion Smarter Models Require More Compute During Inference Tokens = Compute Single Shot Inference PT -4o Reasoning Models GPT-5 Agentic + Reasoning Codex Agentic models orchestrate multiple models, each performing search, reasoning, and synthesis.”
Cerebras fast inference serves as foundation for agentic flows with named customers.
“• Cerebras fast inference serves as the foundation for agentic flows in industries ranging from finance to life sciences with customers including Block, Figma, AlphaSense, and GSK.”
Cerebras describes agentic models that orchestrate multiple models.
“Reasoning Drives the Inference Compute Explosion Smarter Models Require More Compute During Inference Tokens = Compute Single Shot Inference PT -4o Reasoning Models GPT-5 Agentic + Reasoning Codex Agentic models orchestrate multiple models, each performing search, reasoning, and synthesis. Agentic models consume 100 to 1,000 times more tokens than base models. Build a plan Break problem down Solve each part Check each answer Combine parts Draft solution Review solution Illustrative compute required for inference. 1x ~100x >1,000x”
Cerebras describes how agentic models require 100 to 1,000 times more inference compute than base models.
“Reasoning Drives the Inference Compute Explosion Smarter Models Require More Compute During Inference Tokens = Compute Single Shot Inference PT -4o Reasoning Models GPT-5 Agentic + Reasoning Codex Agentic models orchestrate multiple models, each performing search, reasoning, and synthesis. Agentic models consume 100 to 1,000 times more tokens than base models. Build a plan Break problem down Solve each part Check each answer Combine parts Draft solution Review solution Illustrative compute required for inference. 1x ~100x >1,000x”