FINAL-Bench/Darwin-180B-RSI-R3
just landed · new, unvettedmodel
New, unvetted: created or first seen in the last few days, not from a trusted org and not past the signal threshold yet.
Recursive self-improvement (RSI) is the core of this line. Mixture-of-experts, 180B total (512 experts, 10 per token). Context: 256K tokens (config). License: qwen-community-1.0. Vendor-reported: "Darwin is VIDRAFT's measurement-driven reasoning model family: 50+ official models, 400+ community derivatives, and two places in the GPQA Diamond top 3 (Darwin-180B-RSI #1 · Darwin-397B-ZTC #3)." Vendor-reported: "Earlier Darwin generations grafted the best-performing expert/FFN blocks from other strong models onto a base backbone; the RSI line adds a new ingredient: the model's own verified work."
- Vendor-reported: Darwin is VIDRAFT's measurement-driven reasoning model family: 50+ official models, 400+ community derivatives, and two places in the GPQA Diamond top 3 (Darwin-180B-RSI #1 · Darwin-397B-ZTC #3).
- Vendor-reported: Earlier Darwin generations grafted the best-performing expert/FFN blocks from other strong models onto a base backbone; the RSI line adds a new ingredient: the model's own verified work.
- Org
- FINAL-Bench
- Released
- 2026-10-01
- Status
- just landed · new, unvetted
- License
- qwen-community-1.0 (restricted)
- Parameters
- 180.0B
- Architecture
- MoE · 512 experts, 10/token · Qwen4ExpForConditionalGeneration
- Context
- 262,144 tokens
- Task
- image-text-to-text · vision
- Signal
- 38 HF likes · 762 downloads
Runs on
mixed: real files + estimate; need = weights + FP16 KV cache for 8K tokens + 1GB. A sizing calculation, not tested.
- Q4 ~113.1 GB: 1× DGX Spark (128GB), Mac 192GB, Mac 256GB, CPU only, 128GB RAM (slow)
- Q8 ~193.0 GB (estimated size): 2× DGX Spark (256GB), CPU only, 256GB RAM (slow)
- FP16 ~361.8 GB (estimated size): Mac 512GB
Links
First published here 2026-10-01 · updated 2026-10-30 · markdown · JSON