Back home

Endeavor 1.1

x 0.88
Visit source

Intro

Endeavor 1.1 is Flower Labs' frontier generalist model, launched on November 3, 2026, for reasoning, coding and long-horizon agent work. It is Flower's third model release this year (after Lizzy and Endeavor 1.0). Versus 1.0 it strengthens reasoning/coding, adds image inputs, and offers EU-located GPU inference alongside the global managed option.

The pitch is frontier capability with deployment control: use Flower's managed API or self-host in a private cloud / on premises. Readable thinking traces are exposed for audit; outputs and traces belong to the customer and can feed further custom training. Access is by request; the model id is flwrlabs/endeavor-1.1.

Explain

Highlights

  • Performance: Flower claims a composite intelligence score competitive with OpenAI / Anthropic; GPQA rises from 92.0 (1.0) to 94.4, HumanEval to 99.9, with additional internal figures for AIME 2026, IFEval and MMMU-Pro.
  • Multimodal: image inputs land in 1.1.
  • Deployment: Flower-managed (global or EU GPUs), private cloud, or on premises—aimed at regulated teams that need inspectable traces and data ownership.
  • Enterprise: SLAs and deployment support; documented context windows vary by deployment mode.

Lineage

  • Predecessor Lizzy (Apr 2026) was a UK-oriented sovereign 7B; Endeavor 1.0 (~Sep 2026) moved the line to frontier generalist; 1.1 is an upgrade of that product, not a new brand.
  • Distribution is request / private—not casually downloadable open weights on Hugging Face.

Benchmarks and latency are vendor claims; independent third-party replication is still limited.

Usage

Who it is for

  • Regulated / enterprise teams that need self-hosting or EU inference.
  • Agent builders who want readable thinking traces for evals, audit and continual training loops.
  • Buyers tracking Flower's model line (Lizzy → Endeavor).

Getting started

  1. Read flower.ai/blog/2026-11-03-introducing-endeavor-1.1 and flower.ai/models/endeavor.
  2. Request access and specify global vs EU inference.
  3. Call https://api.flower.ai/v1 with model id flwrlabs/endeavor-1.1 on your own evals and agent tasks.
  4. For private deployment, confirm artifact formats, context limits and SLAs during onboarding.

Advertisement