Our Models

Bloom Labs » Our Models

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.

Flagship

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 »
Balanced

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 »
Lightweight

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 »