To access the LLM tools, you (or another review member) will first need to add credit to your user account.
This can be done under the Purchase tab of the Account Manager (at https://eppi.ioe.ac.uk/cms/Default.aspx?tabid=2935).
.
Edit the review's setting.
You can then apply your credit purchase to the review to cover LLM usage.
(If you have multiple credits, just pick the appropriate one.)
Once the credit has been assigned to LLM usage, you will then be able to access the auto-coding tool.
In the References Tab, select the refs to be coded and the relevant coding tool.
(You can choose from a number of LLMs.)
Then Submit the job.
(As well as batch mode, you can work on a single reference.)
Further details of setting up your prompts / codes can be found here.
(Costs for the LLMs currently available in EPPI Reviewer are listed below.)
| LLM |
Description |
Input Cost* |
Output Cost* |
Retirement date |
| OpenAI GPT4 |
This is the first LLM deployed in EPPI Reviewer in 2024. The model used is OpenAI "GPT-4o" with "model version = 2024-08-06". GPT models are optimised for "Chat", making them quicker and more "creative" than reasoning models. |
2 |
8 |
27-Feb-26 |
| GPT 4.1 |
Flagship model of the GPT series. The model used is OpenAI "GPT-4.1" with "model version = 2025-04-14". |
2 |
6 |
11-Apr-26 |
| GPT 4.1 nano |
The model used is OpenAI "GPT-4.1-nano" with "model version = 2025-04-14". This is a quick and cheap model. |
1 |
1 |
11-Apr-26 |
| OpenAI o4-mini |
The model used is OpenAI "o4-mini" with "model version = 2025-04-16". This is a reasoning model, optimised for speed/efficiency. Reasoning models iterate a little with the intention of giving more accurate answers (but evaluation is needed!). This makes them slower than other models. |
1 |
3 |
11-Apr-26 |
| OpenAI o3 |
The model used is OpenAI "o3" with "model version = 2025-04-16". This is the latest OpenAI reasoning model we can currently make available, so it's also slow and expensive. |
8 |
30 |
09-Apr-26 |
| DeepSeek-R1 |
The model used is "DeepSeek-R1". This is a slow reasoning model, which outputs an "answer", preceded by the "reasoning" text. EPPI Reviewer does not retain the Reasoning part of the response. |
1 |
4 |
03-Jul-26 |
| Mistral Large 24.11 |
The model used is "Mistral Large 24.11". This is a multilingual reasoning model, (mostly) open source |
2 |
5 |
11-Apr-26 |
| Llama 3.1 |
The model used is "Llama-3.1-405B-Instruct". This is a multilingual model optimised for assistant-like chat/tasks. |
4 |
12 |
11-Jul-26 |
|
(*Costs are £ per million tokens)