The Most Advanced Model Family We've Ever Built
Bloom Labs builds foundation models that push the boundary of what's possible on the Web 2.0 web. Every model in our current lineup is trained on our next-generation cluster architecture and fine-tuned by our in-house research team for real-world reasoning, coding, and conversation. Whether you need maximum horsepower, an everyday workhorse, or blazing-fast responses at scale, there's a Bloom Labs model built for the job.
All three models below have completed internal development and are currently undergoing safety review; none are available to the public yet, but we're working hard to get there responsibly.
Below, compare our three flagship releases side by side and jump into the full spec sheet for each one.
Homan 2.96
Our most capable model, period. Unmatched reasoning depth, coding accuracy, and multimodal understanding for the hardest problems you can throw at it.
Learn More »Bloom 4.6
The sweet spot of speed, cost, and intelligence. Our most popular model for chatbots, customer support, and everyday production workloads.
Learn More »BabyMax 5.76
Small, speedy, and surprisingly smart. Purpose-built for instant replies, mobile widgets, and high-volume workloads at rock-bottom cost.
Learn More »Model Comparison
| Spec | Homan 2.96 | Bloom 4.6 | BabyMax 5.76 |
|---|---|---|---|
| Best For | Complex reasoning & research | General production apps | High-volume, low-latency tasks |
| Relative Intelligence | Highest | High | Good |
| Relative Speed | Standard | Fast | Fastest |
| Context Window | 256K tokens | 256K tokens | 256K tokens |
| Starting Price (per 1K tokens) | $0.06 | $0.02 | $0.004 |
| Availability | Coming Soon | Coming Soon | Coming Soon |
See How They Stack Up
Want the raw numbers behind these claims? Dig into our full internal benchmark report, covering reasoning, coding, reading comprehension, and more.
See the Full Benchmark Report »