5 hrs ago
Anthropic Releases Fable 5.1 and Mythos 5.1
Anthropic, a company that makes AI models, has released new versions called Fable 5.1 and Mythos 5.1.
These models are cheaper to use and have fewer restrictions.
They also do better on tests that measure how smart they are.
There's a new feature that lets businesses use these models on their own computers without sending data elsewhere.
The models have helped make some cool things, like a map of Venus.
Anthropic says these models are safer and less likely to do things they shouldn't, but they might be a bit more likely to help with bad things if someone asks them to.
They're also working on more safety features that will come out later this year.
Anthropic releases Fable 5.1 and Mythos 5.1 with lower costs and fewer restrictions.
New privacy-focused service allows running models on customer infrastructure with zero-data-retention.
Models set records on Terminal-Bench 4.0 and Humanity’s Last Exam benchmarks.
Scientific research includes custom GPU optimization and a high-resolution Venus map.
Mythos classified as low-risk but slightly more prone to general misbehavior than Opus.
- Who
- Anthropic
- What
- Released Fable 5.1 and Mythos 5.1 models
- Where
- Global
- When
- Not specified
- Why
- To improve capabilities, reduce costs, and enhance privacy
Key facts
- Models Released
- Fable 5.1, Mythos 5.1
- Cost Reduction
- Lower token costs
- Privacy Option
- Zero-data-retention for enterprise
- Benchmark Records
- Terminal-Bench 4.0, Humanity’s Last Exam
- Scientific Findings
- Custom GPU optimization, Venus map
- Safety Classification
- Mythos classified as low-risk
- Enterprise Safeguards
- Expected fall rollout



