OrcaRouter Releases Abliterated DeepSeek V4.1 Flash With 1M Context for Verified Security Researchers
brhydon · x · 2026-10-25
OrcaRouter abliterated DeepSeek V4.1 Flash (1027 checkpoint) itself after requests from the security community, now live on its Cyber Tier alongside OrcaCyber Zero and Ultra, available only to verified security researchers.
Key points:
- Purpose-built for authorized security work: vuln research and reproduction, PoC construction, red-team tooling, code audit, malware/binary analysis, log and traffic triage
- 1M-token context (whole repos, log sets, packet captures in one call), up to 384K output, vision input, function calling, structured JSON
- Reported scores: CyberGym 88.1, SEC-Bench Pro 62.8, Terminal-Bench 2.1 90.6, GPQA Diamond 90.9
- Positioned as the high-throughput, cost-efficient Flash tier for bulk vulnerability triage and agentic security loops
Commercially selling a refusal-removed model is contentious in AI safety circles, making this a rare public productization of that direction.
Related event: OrcaRouter Open-Sources Uncensored DeepSeek V4.1(2 posts)→
More from Models
- Heavy user downgrades from Claude Max to $20 Pro: Opus 5.5 makes it great value — haider1 · 2026-10-25
- Drex 2.0 goes fully local on Apple Silicon: open weights, 113ms median latency — flngr · 2026-10-25
- True-or-false quiz: 5 common misconceptions about LLM tokens and context limits — ClickOk5811 · 2026-10-25
- Leak: OpenAI quietly restarted frontier training Oct 1 with offline web; agents found Brotli RCE — mark_k · 2026-10-25
- Claude weekly usage limit hit 96% as power user begs Anthropic for relief — adonis_singh · 2026-10-25
- ChatGPT voice mode spotted using open-source Silero VAD v6 — ObnoxiouslyVivid · 2026-10-25