Claude Mythos Preview
Anthropic · Launched Apr 2026
Anthropic research preview; strong on deep browsing (BrowseComp 86.9%) and OSWorld-Verified 79.6%.
Benchmark performance
OpenAI's 1,266 hard browsing problems that reward research depth and factual grounding rather than shallow navigation.
Other screen-level os control agents
The 3 agents in this category, ranked by peak benchmark.
| Agent | Maker | Launch | Peak | Pricing |
|---|---|---|---|---|
| Claude Sonnet 4.6 | Anthropic | 2026-02 | 1470.0 | $3 / $15 per M tokens |
| Claude Sonnet 4.5 | Anthropic | 2025-09 | 62.9 | Legacy Anthropic API |
Recent coverage
2026-07-03
CVEs spike 3.5x after Anthropic's Mythos Preview launch
2026-06-26
OpenAI Launches GPT-5.6 Sol Under US Government Restrictions
2026-06-10
Anthropic: Mythos Preview Builds Working Exploits in Hours, Not Weeks
2026-06-04
Anthropic: Claude Authors 80%+ of Code, Task Length Doubling Every 4 Months
2026-05-14
Claude Mythos Clears All UK Cyberattack Simulators, Doubling Speed Revised
2026-05-12
OpenAI Launches Daybreak Cyber Initiative to Rival Anthropic's Glasswing
2026-05-08
Claude Mythos Preview Doubles METR Time Horizon at 80% Success
2026-05-07
Claude Mythos Helped Firefox Fix More Bugs in April Than 15 Prior Months Combined
Quick facts
- Type
- Screen-level OS control
- Maker
- Anthropic
- Launch
- 2026-04-01
- Open source
- No
- Pricing
- Research preview
- Benchmarks scored
- 3
- Article mentions
- 23
- Rank in category
- #2 of 3