{"slug": "show-hn-abliterated-llm-provider-for-cyber-tasks", "title": "Show HN: Abliterated LLM provider for cyber tasks", "summary": "Refuseless launched an abliterated version of GLM 5.3, an open model with refusals removed, served through OpenAI-compatible endpoints with zero prompt retention for cybersecurity teams. The provider reports the model scores 66.9 on DeepSWE agentic coding, 84.5 on CyberGym cybersecurity, and solves 105 of 130 exploits on ExploitGym, and offers $5 in free credits with credit card and cryptocurrency accepted. The lineup includes glm-5.3-abliterated and glm-5.3-flash-abliterated, both with a 1,048,576-token context window and 16,384-token output limit, accessible via the base URL https://api.refuseless.com/v1.", "body_md": "Frontier coding and cybersecurity\n\n### GLM 5.3 Abliterated\n\nThe most powerful open model for cyber tasks to date, with refusals removed.\n\n- DeepSWE**66.9*** agentic coding*\n- CyberGym**84.5*** cybersecurity*\n- ExploitGym**105/130*** exploits solved*\n\nPROMO$5 FREE CREDITS\n\nAbliterated models served via OpenAI-compatible endpoints, zero prompt retention.\n\nCredit card & cryptocurrency accepted\n\nfor cybersecurity teams and professionals\n\nCost-effective cyber intelligence\n\nVery powerful model competing with frontier at much lower cost.\n\nOpenCode and Pi both speak this API.\n\n```\n{\n  \"$schema\": \"https://opencode.ai/config.json\",\n  \"model\": \"refuseless/glm-5.3-abliterated\",\n  \"provider\": {\n    \"refuseless\": {\n        \"npm\": \"@ai-sdk/openai-compatible\",\n        \"name\": \"Refuseless\",\n        \"options\": {\n          \"baseURL\": \"https://api.refuseless.com/v1\",\n          \"apiKey\": \"{env:REFUSELESS_API_KEY}\"\n        },\n        \"models\": {\n          \"glm-5.3-abliterated\": {\n            \"name\": \"GLM 5.3 Abliterated\",\n            \"tools\": true,\n            \"reasoning\": true,\n            \"limit\": {\n              \"context\": 1048576,\n              \"output\": 16384\n            }\n          },\n          \"glm-5.3-flash-abliterated\": {\n            \"name\": \"GLM 5.3 Flash Abliterated\",\n            \"tools\": true,\n            \"reasoning\": true,\n            \"limit\": {\n              \"context\": 1048576,\n              \"output\": 16384\n            }\n          }\n        }\n      }\n  }\n}\n```\n\n`export REFUSELESS_API_KEY=\"rk-…\"` on Mac, `$env:REFUSELESS_API_KEY=\"rk-…\"` on Windows. The sample above reads it via `{env:REFUSELESS_API_KEY}`.`~/.config/opencode/opencode.json`, Windows: `C:\\Users\\<you>\\.config\\opencode\\opencode.json`. Create it if missing. If it already has a `\"provider\"` object, add `\"refuseless\"` inside it — don't overwrite the rest.`opencode --model refuseless/glm-5.3-abliterated`, or `/models` inside the TUI and pick `refuseless / glm-5.3-abliterated`. Model `glm-5.3-abliterated` (context 1,048,576) — all 2 lineup models ship in the block, swap the id after `refuseless/` to switch — send `Say ok.` to verify.\nSame request shape as OpenAI chat completions. Point the base URL at Refuseless, set a lineup id, send the key.\n\n```\ncurl https://api.refuseless.com/v1/chat/completions \\\n  -H \"Authorization: Bearer $REFUSELESS_API_KEY\" \\\n  -H \"Content-Type: application/json\" \\\n  -d '{\n    \"model\": \"glm-5.3-refuseless\",\n    \"messages\": [{\"role\": \"user\", \"content\": \"Write a tight system prompt.\"}]\n  }'\n```\n\nHost and path are the intended contract. Auth is the API key.", "url": "https://wpnews.pro/news/show-hn-abliterated-llm-provider-for-cyber-tasks", "canonical_source": "https://refuseless.com", "published_at": "2026-10-05 12:36:08+00:00", "updated_at": "2026-10-05 12:49:26.583153+00:00", "lang": "en", "topics": ["ai-safety", "large-language-models", "ai-products", "ai-tools", "ai-startups"], "entities": ["Refuseless", "GLM 5.3 Abliterated", "GLM 5.3 Flash Abliterated", "OpenCode", "Pi", "DeepSWE", "CyberGym", "ExploitGym"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/show-hn-abliterated-llm-provider-for-cyber-tasks", "markdown": "https://wpnews.pro/news/show-hn-abliterated-llm-provider-for-cyber-tasks.md", "text": "https://wpnews.pro/news/show-hn-abliterated-llm-provider-for-cyber-tasks.txt", "jsonld": "https://wpnews.pro/news/show-hn-abliterated-llm-provider-for-cyber-tasks.jsonld"}}