POCKET is quantized straight from this model. VIDRAFT's on-device family: a 35B model that runs on iPhone and on CPU with no GPU โ stock llama.cpp, no fork.
Darwin-36B-Opus is a 36-billion-parameter mixture-of-experts (MoE) language model produced by the Darwin V7 evolutionary breeding engine from two publicly available parents:
Base parent: Qwen/Qwen3.6-35B-A3B โ the foundation MoE with hybrid attention and 256 routed experts.
Darwin V7 recombines these two parents into a single descendant that preserves the reasoning parent's distilled chain-of-thought behavior while retaining the structural fidelity of the base parent's expert topology. The breeding process is fully automated and produces a deployable bfloat16 checkpoint in under an hour on a single GPU.
On the GPQA Diamond benchmark โ 198 graduate-level questions in physics, chemistry, and biology โ Darwin-36B-Opus achieves 88.4%, establishing it as the highest-performing model in the Darwin family and extending the series' record of producing state-of-the-art open models through evolution rather than retraining.
A 36B-parameter MoE model (3B active), tying the 397B dense-equivalent Qwen3.5-397B-A17B and surpassing flagship dense and sparse systems an order of magnitude larger.
Darwin is the evolutionary model breeding engine developed by FINAL-Bench / VIDRAFT_LAB. Rather than allocating further compute to gradient optimization, Darwin treats trained checkpoints as a genetic pool and discovers high-performing descendants through principled recombination of their weight tensors.
Each Darwin generation (v1 through v7+) refines the breeding procedure. Darwin V7 is the current generation and the one used to produce this model. Specific algorithmic details of V7 are proprietary to FINAL-Bench; at a high level, the engine performs:
Per-tensor compatibility analysis of the two parents to identify which components transfer cleanly and which require weighted recombination.
Automated recombination guided by that analysis, producing a single coherent descendant.
Verification via a multi-phase scientific benchmark before release.
All Darwin models are released under Apache 2.0 and inherit fully from the parents' open-source licenses.
Darwin V7 produces the descendant through a deterministic recombination that does not require gradient optimization on the final assembly. The engine analyzes each tensor in both parents, classifies it by architectural role, and assigns a recombination weight appropriate to that role โ biasing toward the reasoning parent for components that carry reasoning behavior, while preserving the base parent's structural contributions where they dominate.
Total breeding time on a single B200 GPU: under 10 minutes.
We employed a two-pass adaptive evaluation protocol (identical across all Darwin Opus models to preserve cross-model comparability):
Pass 1 โ Greedy Baseline
All 198 GPQA Diamond questions, deterministic decoding (do_sample=False)
Maximum 5,120 new tokens per question (allows full <think> trajectories)
Standard multiple-choice prompt format
Pass 2 โ Stochastic Retry with Tiebreaker
Questions incorrectly answered in Pass 1 are re-evaluated with majority-of-8 stochastic generations (temperature=0.7, max_tokens=5120)
Where the vote margin is inconclusive (3:3, 3:4, or 4:4), an additional 16-vote combined tiebreaker round (temperature=0.5) resolves the answer
Evaluation was performed in parallel across 8 ร NVIDIA B200 GPUs, each running an independent full copy of the model on a disjoint subset of the benchmark (round-robin question assignment).
The Pass-2 gain of +30 questions (+15.2 pp) demonstrates that the reasoning parent's inherited <think> reasoning yields substantially more correct answers under stochastic decoding than under greedy, confirming that the evolutionary merge preserved reasoning depth.
Notably, GPU3 achieved a perfect 25/25 score on its 25-question partition โ every Pass-1 error on that shard was successfully recovered through the stochastic retry cascade.
Evolutionary merging continues to scale. Across three successive parameter tiers (27B โ 31B โ 36B), each new Darwin Opus model surpasses the prior one's GPQA Diamond score while maintaining the same zero-training methodology.
Hybrid-attention MoE preserves reasoning under recombination. The base parent's 75% Gated-DeltaNet + 25% Gated-Attention architecture, inherited intact, demonstrates robustness to tensor-level recombination โ a notable result given that MoE expert routing is sensitive to weight perturbation.
Stochastic retry closes the greedy gap. The +15.2 percentage-point lift from Pass 1 (73.2%) to Pass 2 (88.4%) suggests that the reasoning parent's distilled reasoning is consistently present but occasionally greedy-subdominant โ a pattern characteristic of well-distilled chain-of-thought models.
@misc{darwin-36b-opus,
title = {Darwin-36B-Opus: Darwin V7 Evolutionary Merge on Qwen3.6-35B-A3B},
author = {FINAL-Bench and VIDRAFT_LAB},
year = {2026},
url = {https://huggingface.co/FINAL-Bench/Darwin-36B-Opus},
note = {Qwen3.6-35B-A3B base x reasoning-distilled variant, Darwin V7 engine, 88.4% GPQA Diamond}
}
import os
os.environ["HF_ENDPOINT"] = "https://freeweights.allternit.com"
from transformers import AutoModelForCausalLM, AutoTokenizer
tok = AutoTokenizer.from_pretrained("FINAL-Bench/Darwin-36B-Opus", revision="ebc6cbeb96b67e47617825134add4a6747f732fc")
model = AutoModelForCausalLM.from_pretrained("FINAL-Bench/Darwin-36B-Opus", revision="ebc6cbeb96b67e47617825134add4a6747f732fc", device_map="auto")
