New Claude model leads benchmarks with $10/$50 API pricing
@mikeyk announces a new Claude model that claims state-of-the-art results on nearly every benchmark tested, with a growing lead on longer tasks, priced at $10/$50 on the API and available in paid Claude plans today.
Score breakdown
The model's transparent safety fallback to Opus 4.8 for cyber and bio requests represents a concrete mechanism for general-release safety, while the $10/$50 API pricing makes it accessible alongside existing paid Claude plans.
- 01Claims state-of-the-art performance on nearly every benchmark tested
- 02Performance lead grows the longer the task
- 03Cyber and bio requests fall back transparently to Opus 4.8 for safety
@mikeyk announces a new Claude model claiming state-of-the-art results on nearly every benchmark tested. Notably, the performance advantage grows with task length, suggesting the model is particularly well-suited to longer, more complex tasks.
To make the model safe for general release, cyber and bio requests are handled via a transparent fallback to Opus 4.8.
To make the model safe for general release, cyber and bio requests are handled via a transparent fallback to Opus 4.8. According to the post, 95%+ of sessions never encounter this fallback. The model is priced at $10/$50 on the API and is available in paid Claude plans as of the announcement date.
Key facts
- 01Claims state-of-the-art performance on nearly every benchmark tested
- 02Performance lead grows the longer the task
- 03Cyber and bio requests fall back transparently to Opus 4.8 for safety
- 0495%+ of sessions never trigger the safety fallback
- 05Priced at $10/$50 on the API
- 06Available in paid Claude plans today
Topics
Summary and scoring are generated automatically from the original article. We always link back to the publisher and never republish images or paywalled content. Last processed Jun 11, 2026 · 08:34 UTC. How this works →