Yesterday I wrote that Anthropic's IPO looked like a last-one-standing situation, with investors split between a $1T and a $1.5T valuation.

Yesterday I wrote that Anthropic's IPO looked like a last-one-standing situation, with investors split between a $1T and a $1.5T valuation. Now the company is shipping models and signing distribution deals as if the IPO debate were happening somewhere else.

Cisco announced at a Melbourne event that the red-teaming tools in its AI Defense suite can call Mythos, Anthropic's bug-finding model, to test customers' AI setups. It comes under existing licenses at no extra cost.

IBM separately said it will integrate Mythos into Concert to find weaknesses traditional scanners miss, figure out which ones pose real risk, and generate fixes for developer approval. In both cases customers get Mythos through the product without getting the model itself.

The deals land a day after Anthropic announced Fable 5.2 and Mythos 5.2. Anthropic says both produce output up to 30% faster than 5.1, at $8 per million input tokens and $40 per million output, with an optional fast mode that's up to 2.5x faster at $16 and $80.

Fable is generally available; Mythos stays restricted to vetted organizations. Anthropic also says it put in its strongest protections yet against distillation, meaning attempts to clone the model without paying the training cost.

Natalie Mead, Anthropic's APAC head of applied AI, says there's a long queue of buyers wanting Mythos access, and that Cisco can help meet that demand. My read: routing demand through Cisco and IBM lets Anthropic serve the queue while keeping the weights in-house. I'm curious whether IPO buyers give that any weight or keep focusing on the compute commitments.

Source:
The Register: https://www.theregister.com/ai-and-ml/2026/11/10/anthropics-super-bug-finding-mythos-model-is-going-into-cisco-and-ibm-products/5308616

Not financial advice. For informational and educational purposes only.