NEW
Meet the new GLiDE family of decision models.
Building decision
models at the frontier
Building decision
models at the frontier
Building decision
models at the frontier
Fastino Labs is an applied research lab building open weight language models and tools for production AI.
Fastino Labs is an applied research lab building open weight language models and tools for production AI.
Fastino Labs is an applied research lab building open weight language models and tools for production AI.
Community love for OUR OPEN MODELS
55M+
55M+
Model downloads
on Hugging Face
Model downloads
on Hugging Face
55M+
Hugging Face model downloads
6.7K+
6.7K+
6.7K+
GitHub
stars
GitHub stars
GitHub
stars
1.1B+
1.1B+
1.1B+
End
users
End users
End
users
NOW AVAILABLE
NOW AVAILABLE
Meet the new Pareto frontier for decision models
Meet the new
Pareto frontier for decision models
Meet the new Pareto frontier for decision models
Five GLiDE decision models from 0.4B to 27B. All open weight.
Five GLiDE decision models from 0.4B to 27B. All open weight.
TRAINING AGENT
An autonomous agent for end-to-end post training
Describe your goal. The Fastino Training Agent handles research, data preparation, training, and evaluation to deliver your own customized open-weight model.
Learn more →
Learn more →
Learn more →
01
Training plan
Researches your task and constraints, selects a base model, and proposes a training plan for your approval.
02
Data engine
Finds, curates, and generates training data. Builds a separate evaluation set before training begins.
03
Experiment
Runs SFT or RL experiments in parallel and uses the results to refine the data and training strategy.
04
Evaluation
Compares candidates against the baseline on held-out data, checks for regressions, and selects the strongest checkpoint.
05
Model delivery
Delivers the selected checkpoint, training data, and a report explaining the experiments and results.
YOUR CODING AGENT
FASTINO MCP CONNECTED
> Use Fastino MCP to post-train a model for financial reasoning over filings and tables. Beat the base on FinQA and TAT-QA.
I’ll check the Fastino catalog and spend, then submit the goal.
Called fastino 3 times (ctrl+o to expand)
Goal submitted. Two decisions only you can make:
· Base models? (Lightning only / any decoder) · Spend ceiling? (cost range comes with the plan)
Lightning only. Show me the cost range first.
Fastino’s plan · nothing trains until you approve
base Nemotron 3.5 Lightning 30B-A3B · method SFT, rank-32 LoRA adapter · data 6 finance skills, held-out split · bar beat base on FinQA and TAT-QA · spend quoted range, your ceiling
› approve
Data engine running…
✓ 6 finance skills built
✓ curated and audited
✓ held-out split frozen · contamination checked
Running experiments…
fastino - get_task_status (MCP)
└ Parallel SFT configurations training
└ Spend ceiling checked before each new run
✓ SFT runs complete
Evaluating candidates…
• champion clears bar
FinQA 15.9 → 59.2
TAT-QA 19.0 → 56.6
ConvFinQA 15.0 → 57.3 unseen
• Report saved: training-report.md ↗
? Promote the champion to a live endpoint? needs your confirmation
› 1. Yes, promote
2. Not yet, keep it in staging
>
▌
~/finance-model
fastino · MCP
Illustrative session. Scores from the published Nemotron 3.5 Lightning Finance run.
CASE STUDY
Nemotron 3.5 Lightning, with NVIDIA ↗
+43.4
FinQA points over base
<10 h
Prompt to two production models
3B
Active parameters, open weights
Models
Models from our lab
Models from our lab
Open-weight language and decision models for structured decisions, information extraction, and domain-specific reasoning. Access models through the Fastino API or download the weights.
Open-weight language and decision models for structured decisions, information extraction, and domain-specific reasoning.
Explore all models →
Explore all models →
Models from our lab
Open-weight language and decision models for structured decisions, extraction, and domain-specific reasoning.
Explore all models →
01
GLiDE
GLiDE
The thinking decision model: structured choices, confidence scores and reasoning for difficult decisions.
The thinking decision model: structured choices, confidence scores and reasoning for difficult decisions.
Decision models
Decision models
02
GLiNER
GLiNER
03
Specialist LLMs
Specialist LLMs
Decision models - GLiDE family
Decision models - GLiDE family
GLiDE
GLiDE
The family of decision models, from 0.4B for the edge to 27B for the hardest choices. Structured decisions with calibrated confidence and reasoning.
The family of decision models, from 0.4B for the edge to 27B for the hardest choices. Structured decisions with calibrated confidence and reasoning.
#1
#1
Fastino GLiDE v1.2
73
73
#1
#1
NACE Drex Max v1.1
73
73
#1
#1
Subquadratic SubQ Snap 1.2
73
73
#2
#2
Baseten Mammoth 27B
71
71
#2
#2
Liquid AI d1-ultra
71
71
#3
#3
Perplexity Decider v1.2
70
70
#4
#4
Baseten Elephant 9B
66
66
#4
#4
TypeSafe Jev 1.14
66
66
Artificial Analysis Decision Index v2.1, November 11, 2026 snapshot.
Artificial Analysis Decision Index v2.1, November 11, 2026 snapshot.
Demo library
Demo library
What will you build?
What will you build?
What will you build?
Demos built with GLiDE and GLiNER for browser agents, robotics, games, simulations, and more.
Demos built with GLiDE and GLiNER for browser agents, robotics, games, simulations, and more.
Demo library →
Demo library →
Demo library →
Inference and post-train open weight models
Inference and post-train language and decision models, all via one API.
Get started
Subscribe to our newsletter
Get the latest updates on model releases, product news, and research.
Fastino Inc. (“Fastino”) develops specialized AI models and provides APIs designed to support structured data extraction, classification, reasoning, and production AI workflows. Fastino is a technology company and does not provide legal, financial, compliance, or advisory services.
Any outputs, predictions, classifications, or decisions generated through Fastino models are based on the configuration, data, and implementation provided by the customer. Fastino does not control, verify, or guarantee the accuracy, completeness, or suitability of model outputs for any specific purpose. By using this website or Fastino’s models and services, you acknowledge that all content and outputs are provided for informational and operational purposes only and agree to our Terms of Use and Privacy Policy.
2026 Fastino Inc.
All rights reserved
Subscribe to our newsletter
Get the latest updates on model releases, product news, and research.
Fastino Inc. (“Fastino”) develops specialized AI models and provides APIs designed to support structured data extraction, classification, reasoning, and production AI workflows. Fastino is a technology company and does not provide legal, financial, compliance, or advisory services.
Any outputs, predictions, classifications, or decisions generated through Fastino models are based on the configuration, data, and implementation provided by the customer. Fastino does not control, verify, or guarantee the accuracy, completeness, or suitability of model outputs for any specific purpose. By using this website or Fastino’s models and services, you acknowledge that all content and outputs are provided for informational and operational purposes only and agree to our Terms of Use and Privacy Policy.
2026 Fastino Inc.
All rights reserved
Subscribe to our newsletter
Get the latest updates on model releases, product news, and research.
Fastino Inc. (“Fastino”) develops specialized AI models and provides APIs designed to support structured data extraction, classification, reasoning, and production AI workflows. Fastino is a technology company and does not provide legal, financial, compliance, or advisory services.
Any outputs, predictions, classifications, or decisions generated through Fastino models are based on the configuration, data, and implementation provided by the customer. Fastino does not control, verify, or guarantee the accuracy, completeness, or suitability of model outputs for any specific purpose. By using this website or Fastino’s models and services, you acknowledge that all content and outputs are provided for informational and operational purposes only and agree to our Terms of Use and Privacy Policy.
2026 Fastino Inc.
All rights reserved