Ultramarine 1 Flash
Open-weights, post-trained model for finance, legal, consulting, and ops knowledge work at API prices built for volume.
Updated 2026-10-15
No free tier is documented for Ultramarine 1 Flash.
Listed pricing: API $0.15/M input, $0.50/M output tokens.
Overview
Ultramarine 1 Flash is a post-trained open-weights model from Paper Instruments aimed at knowledge work in finance, legal, consulting, and operations. It is not a general chat app; teams call it through an API (and soon downloadable weights) when they need domain-heavy reasoning without frontier-model bills.
The pitch is leverage from a tiny lab: a five-person team post-trained GLM-5.3 Flash with roughly $300k of reinforcement learning on synthetic tasks. Public claims put it at or near Claude Fable 5 on their internal suite while costing about 60× less—around $0.13 per task versus $8.19—so the differentiator is cost-performance on professional workflows, not a new consumer UI.
I'd treat it as a specialist workhorse for document-heavy and process-heavy domains where token spend dominates. If you need a polished assistant product, multimodal creation, or a model already baked into an IDE, look elsewhere; if you want open weights plus cheap API inference for FLECO-style work, this is the lane it is competing in.
Is Ultramarine 1 Flash free?
No free tier is documented for Ultramarine 1 Flash.
Listed pricing: API $0.15/M input, $0.50/M output tokens.
No figure is estimated here for what is not published. Check Ultramarine 1 Flash's own site for the current terms.
Pricing on this page has not been re-verified. The entry was last edited on , and no separate pricing check has been run since. Treat the figures as a record of what was published then and confirm on the official site.
Ultramarine 1 Flash pricing
| Plan | Price | What's included |
|---|---|---|
| API | $0.15/M input, $0.50/M output | Live API access for Ultramarine 1 Flash; open weights expected about two weeks after the Oct 14 API launch. Check the site for rate limits and self-host options. |
Live API access for Ultramarine 1 Flash; open weights expected about two weeks after the Oct 14 API launch. Check the site for rate limits and self-host options.
Pricing on this page has not been re-verified. The entry was last edited on , and no separate pricing check has been run since. Treat the figures as a record of what was published then and confirm on the official site.
Is Ultramarine 1 Flash worth it?
There is no free tier to test it on, so the plans above are the entry cost. The recorded trade-offs are listed below, and any one of them can settle the question on its own.
The 8.7/10 AI Score is an editorial read of published capability, price and shipping pace. Nobody here has hands-on hours with Ultramarine 1 Flash. How we verify.
Worth it if
The strengths recorded against this entry.
- Specialized post-training for finance, legal, consulting, and operations knowledge work
- API pricing at $0.15/M input and $0.50/M output keeps high-volume runs cheap versus frontier chat models
- Open-weights path planned soon after API launch for self-host and customization
- Public claims of matching or beating Claude Fable 5 on their suite at ~60× lower cost (~$0.13 vs ~$8.19 per task)
Not worth it if
Any one of these blocks your use case.
- Narrower than general chat models outside FLECO-style professional domains
- Still a new API from a small team—ecosystem, tooling, and long-term reliability are unproven
- Open weights were not same-day as the API; self-hosting depends on the promised release window
- Benchmark wins are on Paper Instruments’ internal suite—results may not transfer to your own evals
Key features
Open Weights
Paper Instruments plans to release model weights shortly after the API launch, so teams can self-host or fine-tune rather than staying locked to a closed endpoint.
Domain Post-Train
Built by RL post-training GLM-5.3 Flash on synthetic finance, legal, consulting, and operations tasks instead of shipping a raw generalist checkpoint unchanged.
Low-Cost API
Listed at $0.15 per million input tokens and $0.50 per million output tokens, positioned for high-volume knowledge-work pipelines where frontier APIs get expensive fast.
Benchmark Focus
Marketed against Claude Fable 5 on an internal task suite with claimed parity or wins at roughly 60× lower cost—useful if your evals match that style of professional work.
How it compares
| Tool | Best for | Pricing | Score |
|---|---|---|---|
| Ultramarine 1 Flash | — | API $0.15/M input, $0.50/M output tokens | 8.7/10 |
| Perplexity AI | Knowledge workers and researchers who want cited, synthesized answers instead of a list of links to click through. | Freemium | 9.4/10 |
| NotebookLM | people who want cited answers and Studio outputs from a specific set of sources. | Free tier; Google AI Pro $19.99/mo | 9.1/10 |
| Inkling-Small | — | Free — open weights download; Tinker Playground access | 8.8/10 |
Compare head-to-head
Comparison explorer Put Ultramarine 1 Flash up against any three tools Opens with Ultramarine 1 Flash already loaded. Add up to three more from the full index and read pricing, features, pros and cons in one table. →Related reading
Grok Voice 3.0's Speech #1 Claim, Audited
Artificial Analysis ranks Grok Voice 3.0 #1 for speech-to-speech. What that score measures, skips, and how minute pricing shifts the claim.
Scion vs Git After Anthropic’s Agent Merge Spike
Anthropic’s merges jumped from 2,500 to 30,000 a day. How Scion, Git, Piper, and Sapling differ for agent-scale monorepos—and what still has no product.
Anthropic’s $16.8B Q3 and Claude’s Price Path
Anthropic’s $16.8B Q3 and adjusted profits reset the Claude pricing debate ahead of a potential November IPO, not another model scoreboard.
Ready to try Ultramarine 1 Flash?
Head to the official site to start with Ultramarine 1 Flash — pricing and plans are listed above.
Visit Ultramarine 1 Flash