Sold by: UkisAI
Deployed on AWS
Swift 1.5 27B is a reasoning model that gets to the answer with far fewer thinking tokens than its base model, without giving up accuracy. Deploy it on Amazon SageMaker AI in minutes.
Overview
Swift 1.5 27B is UkisAI's 27B reasoning model, post-trained from Qwen3.8 to cut reasoning-token usage while keeping accuracy on math, coding and general reasoning. Fewer reasoning tokens means lower latency and lower cost for every request.
The model ships as a GSQ-RCO quantized GGUF served by llama.cpp, so it fits on a single GPU: ml.g7e.2xlarge (RTX PRO 6000, 96 GB) for long context and high concurrency, or ml.g5.2xlarge for smaller workloads. The endpoint exposes an OpenAI-compatible chat completions API with streaming, and returns the model's reasoning separately from its final answer.
Use it for coding assistants, agents, math and analysis, and any workload where reasoning quality matters but token budgets are tight.
Highlights
- Reaches the same answers with far fewer reasoning tokens, so responses are faster and cheaper per query
- OpenAI-compatible chat completions API on /invocations, with streaming and separate reasoning_content
- Runs fully inside your AWS account with network isolation; your prompts and outputs never leave it
Details
Sold by
Categories
Delivery method
Latest version
Deployed on AWS
New
Introducing multi-product solutions
You can now purchase comprehensive solutions tailored to use cases and industries.
Features and programs
Financing for AWS Marketplace purchases
AWS Marketplace now accepts line of credit payments through the PNC Vendor Finance program. This program is available to select AWS customers in the US, excluding NV, NC, ND, TN, & VT.
Pricing
This product is available free of charge. Free subscriptions have no end date and may be canceled any time.
Additional AWS infrastructure costs may apply. Use the AWS Pricing Calculator to estimate your infrastructure costs.
Dimensions summary
The software itself is free, so you pay only for the AWS compute hours used to run the model. Charges follow the instance type you pick and the inference mode. Two instances support real-time inference, where you serve responses on demand: a ml.g7e.2xlarge option and a ml.g5.2xlarge option. The ml.g5.2xlarge also offers a batch mode, which processes grouped requests rather than live traffic. All three dimensions bill per host hour. Your cost scales with how long each instance runs, letting you match spend to your workload and chosen hardware.
Top-of-mind questions for buyers
All three dimensions bill per host hour, so charges accrue only while the instance runs. A stopped instance stops the software metering. Idle time still counts if the instance stays running, since billing follows running host hours, not request volume. Powering the instance off ends the hourly charge.
Real-time mode serves responses on demand, so you keep the instance running to handle live requests. Batch mode processes grouped requests together rather than live traffic. Both bill per host hour. Real-time suits interactive use; batch suits scheduled jobs where you run the instance only during processing.
One host hour is one hour that the chosen instance runs, including its attached GPU and memory. The ml.g7e.2xlarge and ml.g5.2xlarge are separate hardware options you select. You pay the hourly rate for each running hour, regardless of how many requests the model processes in that hour.
ukisai.com
Helpful?
Vendor refund policy
No refunds. You are billed hourly only while your endpoint or batch job runs; delete the endpoint to stop charges. Contact hello@ukisai.com with billing questions.
Legal
Vendor terms and conditions
Upon subscribing to this product, you must acknowledge and agree to the terms and conditions outlined in the vendor's End User License Agreement (EULA) .
Content disclaimer
Vendors are responsible for their product descriptions and other product content. AWS does not warrant that vendors' product descriptions or other product content are accurate, complete, reliable, current, or error-free.
Delivery details
Amazon SageMaker model
An Amazon SageMaker model package is a pre-trained machine learning model ready to use without additional training. Use the model package to create a model on Amazon SageMaker for real-time inference or batch processing. Amazon SageMaker is a fully managed platform for building, training, and deploying machine learning models at scale.
Deploy the model on Amazon SageMaker AI using the following options:
Version release notes
Initial release: Swift 1.5 GSQ-RCO GGUF on llama.cpp with an OpenAI-compatible chat completions API.
Additional details
Inputs
- Summary
OpenAI-style chat completions JSON (real time) or JSON Lines with one request per line (batch).
- Limitations for input type
- Max payload 6 MB per request.
- Input MIME type
- application/json, application/jsonlines
{"messages": [{"role": "user", "content": "What is 17 * 23? Answer with just the number."}], "max_tokens": 512}
{"messages": [{"role": "user", "content": "What is 17 * 23? Answer with just the number."}], "max_tokens": 512}
Input data descriptions
The following table describes supported input data fields for real-time inference and batch transform.
Field name | Description | Constraints | Required |
|---|---|---|---|
messages | Conversation as a list of {role, content} objects. | - | Yes |
max_tokens | Maximum tokens to generate, including reasoning. | - | No |
stream | Return server-sent events when true. | - | No |
Support
Vendor support
Email hello@ukisai.com . We respond within one business day.
AWS infrastructure support
AWS Support is a one-on-one, fast-response support channel that is staffed 24x7x365 with experienced and technical support engineers. The service helps customers of all sizes and technical abilities to successfully utilize the products and features provided by Amazon Web Services.
Similar products
OpenAI GPT-5.5 brings OpenAIs most capable model for complex professional work to coding, analysis, software operation,and long-running agentic tasks through OpenAI APIs and Amazon Bedrock.
Initial release of MedGemma 27B Multimodal (google/medgemma-27b-it) on AWS Marketplace. This version delivers Google's most capable open medical AI model, supporting both medical image comprehension and clinical text reasoning including FHIR-based EHR data, across radiology, dermatology, histopathology, and ophthalmology modalities.
Llama 3.3 Nemotron Super 49B V1.5 is a significantly upgraded version of Llama 3.3 Nemotron Super 49B V1 and is a large language model (LLM) which is a derivative of Meta Llama-3.3-70B-Instruct (AKA the reference model). It is a reasoning model that is post trained for reasoning, human chat preferences, and agentic tasks, such as RAG and tool calling. The model supports a context length of 128K tokens.
By Z.ai
One-stop API access to Z.ai's full suite of GLM models, covering language, vision, image generation, and video capabilities. A single API key unlocks all AI capabilities for intelligent dialogue, content creation, visual understanding, and multimedia generation.
MedGemma 1.5 4B is Google's open-weight multimodal medical AI model, fine-tuned for radiology, pathology, ophthalmology, dermatology, and clinical reasoning. Deploy as a managed SageMaker real-time endpoint in minutes.
Customer reviews
No customer reviews yet
Be the first to review this product . We've partnered with PeerSpot to gather customer feedback. You can share your experience by writing or recording a review, or scheduling a call with a PeerSpot analyst.