Flower launches Endeavor 1.0 for private deployment

Flower Labs launched Endeavor 1.0, a frontier-class generalist available as a managed service or for private deployment, and reports 92.0 on GPQA, 98.2 on HumanEval, 99.9 on AIME 2026, and 94.1 on IFEval in its launch announcement. The company says Endeavor matches GPT-5.6 Sol and Claude Fable 5 on AIME 2026, but access starts as a limited preview and the results are vendor-reported. For teams weighing model control against benchmark quality, this is a deployment option to test against real work—not a benchmark-only switch—alongside the archive’s open-weight adoption analysis.