OpenAI Ultrafast Mode: What 14x Means for Agents (2026)
August 17, 2026
OpenAI's Ultrafast mode runs GPT-5.6 Sol at up to 750 output tokens per second on Cerebras. What the 14x claim means for AI agent latency, and what it omits.
#gpt-56-sol
OpenAI Astra: 10 Math Proofs, Verified in Lean (2026)
August 3, 2026
OpenAI's unreleased Astra model solved ten long-open math problems for about $2,000 and shipped Lean certificates. What agent builders should take from it.
GPT-5.6 Sol: OpenAI's Government-Gated Preview (2026)
June 29, 2026
OpenAI previewed GPT-5.6 Sol, Terra, and Luna on June 26, 2026 — but only to government-vetted partners, via API and Codex, under the cyber EO 14409.