# LLM Training on Llama 70B Emulation 3.3

Source: [LLM Training on Llama 70B Emulation 3.3](<https://www.sota2.com/research/sota/llm-training-on-llama-70b-emulation-3-3>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-02-26T06:06:31.186Z

Evaluation of LLM training efficiency for the Llama 70B Emulation 3.3 model, focusing on the reduction in training time and energy consumption.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | LLM Training |
| Dataset (including any stated split) | Llama 70B Emulation 3.3 |
| Metric | Time Reduction |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 8 of 8 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [Time Reduction](<https://www.sota2.com/research/sota/llm-training-on-llama-70b-emulation-3-3.md?metric=Time%20Reduction>): Higher is better; unit: PERCENT.

- [Energy Reduction](<https://www.sota2.com/research/sota/llm-training-on-llama-70b-emulation-3-3.md?metric=Energy%20Reduction>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Time Reduction | Result date | Report paper |
| --- | --- | --- | --- |
| Kareus | 9.3 | Not recorded | 6978467ca2b28e761d87c477 |
| Kareus | 9.2 | Not recorded | 6978467ca2b28e761d87c477 |
| Kareus | 9.1 | Not recorded | 6978467ca2b28e761d87c477 |
| Kareus | 9.1 | Not recorded | 6978467ca2b28e761d87c477 |
| Megatron-LM + Perseus | 0 | Not recorded | 6978467ca2b28e761d87c477 |
| Megatron-LM + Perseus | 0 | Not recorded | 6978467ca2b28e761d87c477 |
| Megatron-LM + Perseus | 0 | Not recorded | 6978467ca2b28e761d87c477 |
| Megatron-LM + Perseus | 0 | Not recorded | 6978467ca2b28e761d87c477 |

### Kareus

Time Reduction: 9.3. Result ID: 69991afb5698665ec4818d3b.

[Reporting paper](<https://www.sota2.com/research/paper/6978467ca2b28e761d87c477.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6978467ca2b28e761d87c477.md>)

Recorded evaluation configuration: {"# Microbatches":"16","Throughput Configuration":"max-throughput"}. This record does not assert independent reproduction.

### Kareus

Time Reduction: 9.2. Result ID: 69991afb5698665ec4818d3d.

[Reporting paper](<https://www.sota2.com/research/paper/6978467ca2b28e761d87c477.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6978467ca2b28e761d87c477.md>)

Recorded evaluation configuration: {"# Microbatches":"32","Throughput Configuration":"max-throughput"}. This record does not assert independent reproduction.

### Kareus

Time Reduction: 9.1. Result ID: 69991afb5698665ec4818d3f.

[Reporting paper](<https://www.sota2.com/research/paper/6978467ca2b28e761d87c477.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6978467ca2b28e761d87c477.md>)

Recorded evaluation configuration: {"# Microbatches":"64","Throughput Configuration":"max-throughput"}. This record does not assert independent reproduction.

### Kareus

Time Reduction: 9.1. Result ID: 69991afb5698665ec4818d41.

[Reporting paper](<https://www.sota2.com/research/paper/6978467ca2b28e761d87c477.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6978467ca2b28e761d87c477.md>)

Recorded evaluation configuration: {"# Microbatches":"128","Throughput Configuration":"max-throughput"}. This record does not assert independent reproduction.

### Megatron-LM + Perseus

Time Reduction: 0. Result ID: 69991afb5698665ec4818d3a.

[Reporting paper](<https://www.sota2.com/research/paper/6978467ca2b28e761d87c477.md>)

Recorded evaluation configuration: {"# Microbatches":"16","Throughput Configuration":"max-throughput"}. This record does not assert independent reproduction.

### Megatron-LM + Perseus

Time Reduction: 0. Result ID: 69991afb5698665ec4818d3c.

[Reporting paper](<https://www.sota2.com/research/paper/6978467ca2b28e761d87c477.md>)

Recorded evaluation configuration: {"# Microbatches":"32","Throughput Configuration":"max-throughput"}. This record does not assert independent reproduction.

### Megatron-LM + Perseus

Time Reduction: 0. Result ID: 69991afb5698665ec4818d3e.

[Reporting paper](<https://www.sota2.com/research/paper/6978467ca2b28e761d87c477.md>)

Recorded evaluation configuration: {"# Microbatches":"64","Throughput Configuration":"max-throughput"}. This record does not assert independent reproduction.

### Megatron-LM + Perseus

Time Reduction: 0. Result ID: 69991afb5698665ec4818d40.

[Reporting paper](<https://www.sota2.com/research/paper/6978467ca2b28e761d87c477.md>)

Recorded evaluation configuration: {"# Microbatches":"128","Throughput Configuration":"max-throughput"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.