Every few months, one announcement reshapes the entire AI landscape. Leaked signals suggest Anthropic may be about to fire that shot.
The AI model war has a new contender stepping into the ring. Unverified but credible-sounding leaks circulating in developer and AI research communities suggest that Anthropic’s Claude Mythos — the company’s most advanced frontier model, previously available only to a small number of trusted organisations — may be on the verge of a broad public release. And if the rumoured pricing holds up, it’s going to put serious pressure on every competitor in the premium AI segment.
Note: The pricing and timing details in this article are based on unconfirmed leaks and community reports, not official Anthropic announcements. Treat them accordingly.
Table of Contents
What the Leaks Suggest
| Detail | Leaked / Rumoured Spec |
|---|---|
| Model Name | Claude Mythos (Preview → Full Release) |
| Input Token Price | $16 per 1 million tokens |
| Output Token Price | $80 per 1 million tokens |
| Release Timing | Imminent — possibly same week as GPT-5.6 |
| Current Status | Limited access via Project Glasswing |
| Official Confirmation | None at time of publication |
The pricing, if accurate, positions Mythos firmly at the ultra-premium end of the frontier model market — above Claude Opus 4.6’s current pricing tier, and in the same bracket as the most capable models from OpenAI and Google. This is not a mass-market model. This is Anthropic’s answer to the question: what does the absolute ceiling of AI capability look like right now?
What Is Claude Mythos — And What Is Project Glasswing?
For readers unfamiliar with the backstory: Claude Mythos is not a new name. Anthropic has been quietly operating it under Project Glasswing — a restricted early-access programme that has made the model available to a small number of vetted organisations, primarily due to cybersecurity concerns around its capabilities at the frontier level.
The fact that it has been held back from public release at all is itself a signal. Anthropic has been unusually cautious about deploying Mythos broadly, suggesting the model operates in capability territory that required additional safety evaluation before wide rollout. That caution, paradoxically, is exactly what makes the rumoured imminent release so significant — it implies Anthropic believes it has cleared those bars.
More details on Project Glasswing are available at anthropic.com/glasswing.
The Timing: Why a GPT-5.6 Simultaneous Launch Would Matter
The leak suggests Mythos could drop in the same week as OpenAI’s GPT-5.6 — and if that timing is intentional, it’s a bold strategic move.
Simultaneous frontier releases from the two leading AI safety-focused labs would force developers, enterprises, and the broader tech press to make direct comparisons in real time. Benchmarks would be run side by side. API pricing would be weighed against capability. Developer communities would split into camps overnight.
For Anthropic — which has historically operated on a slightly quieter cadence than OpenAI’s aggressive release schedule — launching in the same window as a GPT update would be a clear signal that the company is done ceding the news cycle.
OpenAI’s current model lineup and pricing is publicly tracked at platform.openai.com/docs/models for anyone wanting to run the direct comparison once Mythos pricing is confirmed.
What About Gemini 3.5 Pro?
The leak’s framing explicitly calls out Google’s Gemini 3.5 Pro as the model facing the most pressure — and that framing makes sense contextually. Gemini 3.5 Pro currently occupies the premium multimodal frontier position that both OpenAI and Anthropic are targeting with their top-tier releases.
A Claude Mythos public launch at $16/$80 per million tokens, if matched by strong benchmark performance, would hand enterprise customers a direct alternative to Gemini at the highest capability tier — accelerating the three-way race between Google, OpenAI, and Anthropic that has defined the AI industry for the past two years.
For our previous coverage of where each of these models currently stands, check out our AI model comparison coverage on TechnoSports.
Should Developers Pay Attention Right Now?
Yes — with appropriate scepticism. Here’s the practical checklist:
- Watch Anthropic’s official channels — anthropic.com and @AnthropicAI on X — for any official announcement
- Don’t build production pricing assumptions around the leaked $16/$80 figures until confirmed
- Do start evaluating whether a frontier-level model at this price point fits your use case — if the leaks are directionally correct, the release window is short
We’ll update this article the moment official pricing and availability are confirmed by Anthropic.
⚠️ Disclaimer: All pricing, timing, and release details in this article are based on unverified community leaks and have not been officially confirmed by Anthropic. This article does not constitute financial or investment advice. TechnoSports will publish a full breakdown upon official announcement.
