Your AWS account
- Training runs on GPUs in your own account, and AWS bills you directly at your rates.
- Your data, checkpoints and models never leave your S3.
- Best for customer records, health or financial data.
No SynqTune fee in early access
Replace an LLM API call with a model you own.
Trained on your examples, and tested against the AI you use today.
Or start free: look around with sample runs first, then train on 100 free credits with a work email. No AWS account needed.
- Platform fee
- $0 in early access
- Runs in
- Your AWS or SynqTune GPUs
- Free to start
- 100 credits with a work email
- Switch when
- The report says ready
A 2B model, against the APIs teams pay for
We trained a small open-source model on four public datasets and tested it against the AI APIs companies pay for, on examples neither had seen. Every test is written up, including where the small model loses.
Routing consumer complaints
Consumer finance complaints sorted into 9 categories, 200 test cases
83.5%
Our small model (Gemma 2B)
82.0%
Grok 4 Fast
Level on accuracy, and it answered in 75 milliseconds instead of 5.6 seconds.
Read the write-upLabelling contract clauses
Contract clauses labelled by type, 249 test cases
92.4%
Our small model (Gemma 2B)
90.8%
Gemini 3.5 Flash-Lite
Slightly more accurate, at less than half the cost per clause.
Read the write-upReading receipts into JSON
Shop receipts turned into structured data, 185 test cases
70.3%
Our small model (Gemma 2B)
63.2%
Gemini 3.5 Flash-Lite
More receipts read fully right, and even its slow answers took 3.4 seconds against 15.3.
Read the write-upDetecting banking intents
Bank customer messages sorted into 77 intents, 270 test cases
85.2%
Our small model (Gemma 2B)
78.2%
Gemini 3.5 Flash-Lite
7 points ahead, a lead that holds up statistically.
Read the write-up
Your own report runs the same comparison on your data, in your account. See a real report
What SynqTune does
Companies pay AI providers for millions of small, repetitive tasks, like sorting support tickets. SynqTune trains a small AI model on your own examples, in your AWS account or on SynqTune GPUs, then tests it against the AI you pay for today. If it matches, you can run your own cheaper, private model instead.
For engineers
SynqTune fine-tunes small open models with LoRA on your labeled data, on GPUs in your own AWS account, starting from task templates. It then runs a parity report against your current LLM on held-out examples and gives a clear verdict. Your data and weights stay in your S3. No ML team needed.
Run it in your AWS, or on ours
Pick one per organization. You can start on SynqTune GPUs and move to your own AWS account later.
SynqTune GPUs
- No AWS account needed: runs use GPUs in SynqTune's own AWS account.
- Start with 100 free credits when you sign up with a work email, and ask for more any time.
- Each organization gets its own private storage; move to your own AWS whenever you want.
Buying credits coming soon · $1.50 per 100
From a task to a model you own
Six steps. Everything after the first happens in your AWS account.
Pick a task
Sorting, pulling out fields, tagging, hiding personal details or summarizing. Tell it your categories or the fields you need.
Bring examples
A file you already have, an upload, or the requests you already send to your AI provider.
Clean and check
Duplicates, bad answers and personal details are cleaned out, and you can preview every step first.
Fine-tune
A small open-source model learns your task on a GPU in your AWS account.
Compare
Both models answer the same test examples. The report says whether yours is ready.
Switch
Download your model and run it wherever you want. It is yours to keep.
Pick a task, get a pipeline
Each task is a template: fill in your labels or schema, point it at your examples, and it lays out the steps for you.
Classification and routing
One label per input, at the volume your queue actually has.
Where teams use it
- Support ticket routing
- Intent detection for assistants and phone menus
- Triage and priority
- Content moderation categories
What the template sets up
- Your examples
- Remove duplicates
- Fine-tune
- Compare with your AI
- Save model
How the switch is checked
Your new model and the AI you use today answer the same test examples, and the report compares them before you decide.
In a regulated industry? Every task runs in your own AWS account. How data stays in your account
SynqTune coordinates the work. Your account does it.
Your data and your model stay in your AWS account. SynqTune only starts and stops the work, through a single permission you can remove at any time, and every permission it has is published.
Compute in your account
Training runs on servers in your own AWS account, shut down as soon as the job is done.
Files in your storage
Your data and the trained model are saved only to your own AWS storage.
One permission, revocable
SynqTune gets a single, narrow permission you create. It can't see your saved API keys, and you can remove it at any time.
Common questions
More on setup, every pipeline step and troubleshooting is in the documentation.
Why would a small model be as good as a big one?
Because it only does one narrow job. A large AI model is a generalist; a small model trained on thousands of your own examples becomes a specialist at your task, such as routing your tickets into your categories. It won't write essays, but on that one task it can match the big model, and the comparison report shows whether it does.
Can I use my model in production today?
Training and testing are ready today, free during early access. You can download your model in standard open formats and run it yourself, or chat with it on a server in your account. A hosted endpoint your applications call directly is next on the roadmap, and early-access teams help shape it.
How do I know the small model is good enough to switch?
Turn on the comparison: the AI you use today, with the prompt you use in production, answers the same test examples as your new model. The report scores both for your task, shows where the small model is weaker, and ends with a clear verdict: ready to replace, close, or not ready.
Why train a small model instead of using a big AI model or RAG?
They solve different problems. A big AI model is best for open-ended work, and RAG (retrieval) for answering from documents that change. For a narrow, repeated task with examples of good answers, a small trained model learns the task itself, so each request is short, costs server time instead of per-token fees, and stays in your AWS account. They also combine: retrieval can feed facts to your own model.
Does my data leave my AWS account?
Your training data, your model and your test examples stay in your own AWS storage. SynqTune keeps only your settings, run status and the scores. If you turn on grading with an outside AI or the comparison with your current one, those records go to the provider you choose.
Do I need ML or DevOps experience?
No. Setup is a ready-made template you launch once from the AWS console, and every task starts from a template with sensible defaults. Every setting is visible and explained before you launch.
We'll run your first comparison with you
Bring one task you pay an LLM API for. We set it up together, in your AWS account, and you see the report on your own data before deciding anything.
A 30-minute call
Tell us the task you'd move off the API. We set up the run with you, on the call.
Run in your AWS account
Your examples never leave it. You get the comparison report on your own data, free.
You shape what we build
Early partners' requests go to the top of the roadmap, and launch pricing is locked in.
Prefer email? support@synqtune.com