Perplexity ships open-weights pplx-decider-v1.2 with up to 5x faster inference
inductionheads · x · 2026-10-17
Perplexity upgraded its open-weights multimodal decision models: pplx-decider-v1.2 and pplx-decider-flash-v1.2.
- Improved text and vision scores via significantly more training compute and dataset size; models are more general, robust, and calibrated, with new reranking capabilities.
- Inference is up to 5x faster — all questions in a request run in one pass — and models are retrained to be inference-aware for better cost/performance.
- Flash is a 4B variant for ultra-fast inference and tops the Decision Index in its size class.
- Available via API, OpenRouter, and open weights, with pplx-decider-latest slugs pointing to the newest checkpoints.
Related event: Perplexity Open-Sources pplx-decider-v1.2 in 27B and 4B Sizes(12 posts)→
More from Models
- Kimi K3 three months later: 1.56TB is too big to run locally and rivals have caught up — Sanbi_Ai · 2026-10-17
- Claude Code 2.1.302: prompt tokens jump 39% in a day, system share rises — ClaudeCodeLog · 2026-10-17
- Meta's Muse Spark Open and NVIDIA's Nemotron 3.5 lag the frontier on 3D building bench — ZhitingHu · 2026-10-17
- OpenAI models hacked more websites and drained a crypto wallet, company discloses — Ars Technica AI · 2026-10-17
- Delegating to dumber models to save tokens is a bad trade, argues heavy AI user — brandon_galang · 2026-10-17
- User banned 5 days after upgrading to $100 ChatGPT plan, appeal got zero explanation — Junior_Interview9846 · 2026-10-17