MISSINGNO. A stealth coding and agentic reasoning model on Nous Portal, usable today in Hermes Agent. Nous presents it as highly token-efficient and free for a limited time, without naming the underlying lab. Independent testers report roughly 256K–600K context, tool calling, adjustable reasoning effort, a Qwen-matching tokenizer, and native OpenAI gpt-oss Harmony formatting.
Visit productFeatured media
Loading media…
View on XCosigns Received
“Running MISSINGNO on my hermes agent and it is super fast, really great model!”
“This model (stealth/missingno) looks like "too good to be true" My humble tests result (accessed via @NousResearch api): Verdict: strong agentic candidate, follows tool protocol, won't invent data to please, holds multiple constraints at once. - Checked against that verdict, still green. Looks like a very good model.”
