{"version":1,"type":"story","url":"https://digestai.news/story/cantina-security-releases-apex-flash-1-a-321-3b-open-model-for-vulnera","json":"https://digestai.news/story/cantina-security-releases-apex-flash-1-a-321-3b-open-model-for-vulnera.json","markdown":"https://digestai.news/story/cantina-security-releases-apex-flash-1-a-321-3b-open-model-for-vulnera.md","slug":"cantina-security-releases-apex-flash-1-a-321-3b-open-model-for-vulnera","headline":"Cantina Security releases apex-flash-1, a 321.3B open model for vulnerability research","summary":"Cantina Security, together with Yeta Labs, announced apex‑flash‑1, an open‑weights model built on Z.ai’s GLM‑5.3‑Flash and released on Hugging Face under an MIT license. The 321.3 B‑parameter model was fine‑tuned with GRPO, using a rank‑256 LoRA plus selective full‑parameter training on 150 tasks derived from 50 real vulnerability cases.\n\nIn Cantina’s internal benchmark of 60 held‑out tasks, apex‑flash‑1 solved 40 of them, achieving a 66.7 % pass@1 rate at an estimated cost of $2.38 per run. By comparison, Claude Opus 5 High solved 43 tasks (71.7 % pass@1) but cost $74.68, roughly $0.06 per solved task for apex‑flash‑1 versus $1.74 for Opus. The model runs on vLLM, SGLang or Transformers, but BF16 inference requires about 640 GB of GPU memory. Cantina positions apex‑flash‑1 as a worker model that can be orchestrated by larger systems, aiming to give defenders a locally controllable security‑focused AI.","keyPoints":["Cantina Security and Yeta Labs released apex‑flash‑1, a 321.3B open‑weights security model under MIT license.","In a 60‑task benchmark, apex‑flash‑1 solved 40 tasks (66.7% pass@1) at about $2.38, versus Claude Opus 5 High’s 71.7% at $74.68.","The model runs on vLLM, SGLang or Transformers but needs roughly 640 GB GPU memory for BF16 inference."],"whyItMatters":"An open, locally runnable security model lowers the cost and dependency on proprietary AI services, expanding accessible vulnerability research.","category":{"slug":"models","name":"Generative AI & Models","url":"https://digestai.news/category/models"},"entities":{"companies":["Cantina Security","Yeta Labs","Z.ai","Hugging Face","Anthropic"],"models":["apex-flash-1","GLM-5.3-Flash","Claude Opus 5 High"],"people":[]},"firstPublishedAt":"2026-10-05T01:47:28Z","updatedAt":"2026-10-05T01:47:28Z","sourceCount":1,"hasPrimarySource":false,"sources":[{"outlet":"MarkTechPost","title":"Can an Open Model Do Security Research? Cantina’s apex-flash-1 Solves 40 of 60 Held-Out Bug Tasks","url":"https://marktechpost.com/2026/10/04/can-an-open-model-do-security-research-cantinas-apex-flash-1-solves-40-of-60-held-out-bug-tasks","publishedAt":"2026-10-05T01:47:28Z","type":"press","primary":false,"lead":true}],"sourceNotes":null,"discussions":[],"thread":null,"cite":{"text":"Digest AI, \"Cantina Security releases apex-flash-1, a 321.3B open model for vulnerability research\", 5 October 2026, https://digestai.news/story/cantina-security-releases-apex-flash-1-a-321-3b-open-model-for-vulnera","publisher":"Digest AI","title":"Cantina Security releases apex-flash-1, a 321.3B open model for vulnerability research","datePublished":"2026-10-05T01:47:28Z","url":"https://digestai.news/story/cantina-security-releases-apex-flash-1-a-321-3b-open-model-for-vulnera"},"generatedBy":"Written by Digest AI's editorial model from the linked sources; the sources are the record.","license":"Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse"}