Ripple-65B Preview

Available

Ricursive IntelligenceOpen weights

Ricursive Intelligence's first public model, released Nov 4 2026: a 65B-parameter dense Transformer pretrained from scratch on 12T predominantly English and code tokens, with up to 1M-token context. Text-only. A general-purpose variant of the proprietary model behind Ricursive's chip design platform, lightly post-trained with a focus on agentic coding. Vendor-reported 80.52 on TerminalBench 2.1. Open weights on Hugging Face (BF16 instruct, NVFP4 instruct, BF16 base, plus a DFlash2 speculative-decoding drafter); also served on Together AI, GMI Cloud and OpenRouter.

Official page ↗Model card ↗

Specifications

License: Open weights
Weights: Downloadable
Architecture: Dense
Parameters: 65B
Context window: 1.0M tokens
Max output: —
Knowledge cutoff: —
Price (in / out, $/M): —
Modalities: TextCode

Benchmarks

No benchmark scores recorded yet. Spotted some? Submit a correction.

Vendor-reported figures are claims until independently verified. See methodology.