A large LLM created by combining two fine-tuned Llama 70B models into one 120B model. Combines Xwin and Euryale. Credits to - [@chargoddard](https://huggingface.co/chargoddard) for developing the framework used to merge the model - [mergek…
Model details →Goliath 120B vs Sao10K: Llama 3.1 70B Hanami x1
Sao10K: Llama 3.1 70B Hanami x1 wins on 3 of 6 axes — pricing and capability skew in its favour for most workloads.
This is [Sao10K](/sao10k)'s experiment over [Euryale v2.2](/sao10k/l3.1-euryale-70b).
Model details →Side-by-side comparison
| Capability | Goliath 120B | Sao10K: Llama 3.1 70B Hanami x1 | Winner |
|---|---|---|---|
| Context window Maximum number of input tokens the model can attend to in a single request. | 6K | 16K | 🏆 Sao10K: Llama 3.1 70B Hanami x1 |
| Input price (per 1M) Cost per million input tokens billed by the provider. | $3.75 | $3.00 | 🏆 Sao10K: Llama 3.1 70B Hanami x1 |
| Output price (per 1M) Cost per million output tokens billed by the provider. | $7.50 | $3.00 | 🏆 Sao10K: Llama 3.1 70B Hanami x1 |
| Tool / function calling First-class support for emitting structured tool calls. | — | — | Tie |
| Vision input Accepts image inputs alongside text. | — | — | Tie |
| Reasoning mode Internal chain-of-thought / extended-thinking support. | — | — | Tie |
Frequently asked questions
Is Goliath 120B better than Sao10K: Llama 3.1 70B Hanami x1?
Sao10K: Llama 3.1 70B Hanami x1 wins on 3 of 6 axes — pricing and capability skew in its favour for most workloads.
What's the price difference between Goliath 120B and Sao10K: Llama 3.1 70B Hanami x1?
Input: $3.75 vs $3.00 per 1M tokens. Output: $7.50 vs $3.00 per 1M tokens.
What context windows do Goliath 120B and Sao10K: Llama 3.1 70B Hanami x1 support?
Goliath 120B supports up to 6K tokens. Sao10K: Llama 3.1 70B Hanami x1 supports up to 16K tokens.
Do both Goliath 120B and Sao10K: Llama 3.1 70B Hanami x1 support tool calling?
Goliath 120B: not advertised. Sao10K: Llama 3.1 70B Hanami x1: not advertised.