Ultramarine 1 Flash

Ultramarine 1 Flash

Paper Instruments
+
+

Related Products

  • Google AI Studio
    41 Ratings
    Visit Website
  • Gemini Enterprise Agent Platform
    1,161 Ratings
    Visit Website
  • LM-Kit.NET
    29 Ratings
    Visit Website
  • Cloverleaf
    189 Ratings
    Visit Website
  • Interfacing Integrated Management System (IMS)
    66 Ratings
    Visit Website
  • Innoslate
    98 Ratings
    Visit Website
  • RealEstateAPI (REAPI)
    58 Ratings
    Visit Website
  • AI Video Cut
    1 Rating
    Visit Website
  • Elation Health
    90 Ratings
    Visit Website
  • VKS
    26 Ratings
    Visit Website

About

​Reka Flash 3 is a 21-billion-parameter multimodal AI model developed by Reka AI, designed to excel in general chat, coding, instruction following, and function calling. It processes and reasons with text, images, video, and audio inputs, offering a compact, general-purpose solution for various applications. Trained from scratch on diverse datasets, including publicly accessible and synthetic data, Reka Flash 3 underwent instruction tuning on curated, high-quality data to optimize performance. The final training stage involved reinforcement learning using REINFORCE Leave One-Out (RLOO) with both model-based and rule-based rewards, enhancing its reasoning capabilities. With a context length of 32,000 tokens, Reka Flash 3 performs competitively with proprietary models like OpenAI's o1-mini, making it suitable for low-latency or on-device deployments. The model's full precision requires 39GB (fp16), but it can be compressed to as small as 11GB using 4-bit quantization.

About

Ultramarine 1 Flash is a hosted large language model offered through an API for developers and organizations building AI-assisted professional workflows. It is designed for knowledge-work tasks across areas such as finance, law, consulting, and operations, including research, analysis, and preparation of document-based deliverables. Applications can send prompts and receive streamed or standard responses, use reasoning controls and custom tool calls, and request structured output. The API accepts image inputs and supports OpenAI-compatible chat-completion and Responses formats, as well as Anthropic Messages-compatible requests. Developers can connect through the OpenAI or Anthropic SDKs, Claude Agent SDK, or Claude Code, and can use the model with Paper Instruments' Word, PowerPoint, and Excel libraries. Usage is billed by tokens against a prepaid balance.

Platforms Supported

Windows Not Supported
Mac Not Supported
Linux Not Supported
Cloud Supported
On-Premises Not Supported
iPhone Not Supported
iPad Not Supported
Android Not Supported
Chromebook Not Supported

Platforms Supported

Windows Not Supported
Mac Not Supported
Linux Not Supported
Cloud Supported
On-Premises Not Supported
iPhone Not Supported
iPad Not Supported
Android Not Supported
Chromebook Not Supported

Audience

Developers seeking an AI model for coding assistance, natural language understanding, and multimodal data processing in their applications

Audience

Developers and organizations building AI-enabled knowledge-work tools in finance, law, consulting, and operations.

Support

Phone Support Not Supported
24/7 Live Support Not Supported
Online Supported

Support

Phone Support Not Supported
24/7 Live Support Not Supported
Online Not Supported

API

Offers API Not Supported

API

Offers API Supported

Screenshots and Videos

Screenshots and Videos

Pricing

No information available.
Free Version Not Supported
Free Trial Not Supported

Pricing

$0.15/1M input tokens
Usage is prepaid and billed per token: input costs $0.15 per 1M tokens, cached input costs $0.03 per 1M tokens, and output, including reasoning, costs $0.50 per 1M tokens.
Free Version Not Supported
Free Trial Not Supported

Reviews/Ratings

Overall 0.0 / 5
ease 0.0 / 5
features 0.0 / 5
design 0.0 / 5
support 0.0 / 5

This software hasn't been reviewed yet. Be the first to provide a review:

Review this Software

Reviews/Ratings

Overall 5.0 / 5
ease 5.0 / 5
features 5.0 / 5

Pros & Cons from Real Users

Pros

  • It's so good at professional work that does not have one obvious “correct” answer. It is especially strong for things like research, financial analysis, diligence, writing, business operations, and turning messy information into a polished deliverable. The cost-performance ratio is probably the most impressive part. Paper Instruments reports an average cost of just $0.13 per task on its internal knowledge-work benchmark, while scoring 59.3 compared with 57.8 for Claude Fable 5 and 55.1 for GPT-5.6 Sol.

Cons

  • The main downside is that coding is not really its biggest strength. Its Terminal-Bench and DeepSWE scores are competitive, but models like GPT-5.6 Sol, Gemini 3.8 Flash, and DeepSeek V4.1 Flash still score higher on those coding-focused benchmarks.

Training

Documentation Supported
Webinars Not Supported
Live Online Not Supported
In Person Supported

Training

Documentation Supported
Webinars Not Supported
Live Online Not Supported
In Person Not Supported

Company Information

Reka
Founded: 2022
United States
www.reka.ai/news/introducing-reka-flash

Company Information

Paper Instruments
Founded: 2025
United States
www.paperinstruments.com/ultramarine

Alternatives

Smaug Flash

Smaug Flash

Abacus.AI

Alternatives

MiMo-V2.6-Flash

MiMo-V2.6-Flash

Xiaomi Technology
GLM-4.1V

GLM-4.1V

Z.ai

Categories

AI Coding Models Supported
AI Models Supported

Categories

AI Models Supported

Integrations

Claude Agent SDK Not Supported
Claude Code Not Supported
Nexus Supported
OpenAI Not Supported
OpenAI Agents SDK Not Supported
Space Supported

Integrations

Claude Agent SDK Supported
Claude Code Supported
Nexus Not Supported
OpenAI Supported
OpenAI Agents SDK Supported
Space Not Supported
Claim Reka Flash 3 and update features and information
Claim Reka Flash 3 and update features and information
Claim Ultramarine 1 Flash and update features and information
Claim Ultramarine 1 Flash and update features and information