Ultramarine 1 Flash released
Paper Instruments · Model release
Source confirmed
Data as of October 14, 2026 · How the score is built
CurrentProprietaryReasoning
Released Oct 14, 2026Paper Instruments Ultramarine technical report
Ultramarine 1 Flash
Decision readingUltramarine 1 Flash is tracked, but not publicly ranked yet. The profile exposes 5 sourced benchmark rows and leaves unsupported fields blank until a published record exists.
Released Oct 14, 2026 — see all recent releases
Ultramarine 1 Flash will be repriced, updated or retired. Get each notice with its source and date. Follow model changes
Decision snapshot
Each value carries a field reference instead of floating alone. Markers compare this model with the current ranked and priced catalog; they are not absolute quality thresholds.
Capability
Unranked
field median 50Not eligible for a public rank
Price
$0.15input / $0.50 output
input median $0.78cached $0.030 · blended $0.33
Speed
Not measured
field median 86 tok/sTime to first token not measured
Context
Not published
field median 256,000No exact-model context limit is sourced
Strongest published evidence
Published rows are visible, but no category has enough eligible evidence for a comparative rank.
Validate before choosing
5 published rows leave some tracked benchmark slots empty. Independent runtime speed has not been measured.
Source-linked · 5 displayable benchmark rows
Follow model changesCategory score record
Scores and ranks appear only where published evidence can be displayed. The table keeps the score, weight, cohort, and evidence state together.
| Category | Score | Rank | Percentile | Weight | Benchmarks | Evidence |
|---|---|---|---|---|---|---|
| AgenticRank Not rankedWeight 22%3 benchmarksVerified | 80.3 | 3 benchmarks | Verified | |||
| CodingWeight 20%2 benchmarksVerified | Score pending | 2 benchmarks | Verified | |||
| ReasoningWeight 17%0 benchmarksNot measured | Not measured | 0 benchmarks | Not measured | |||
| MultimodalWeight 12%0 benchmarksNot measured | Not measured | 0 benchmarks | Not measured | |||
| KnowledgeWeight 12%0 benchmarksNot measured | Not measured | 0 benchmarks | Not measured | |||
| MultilingualWeight 7%0 benchmarksNot measured | Not measured | 0 benchmarks | Not measured | |||
| Inst. FollowingWeight 5%0 benchmarksNot measured | Not measured | 0 benchmarks | Not measured | |||
| MathWeight 5%0 benchmarksNot measured | Not measured | 0 benchmarks | Not measured |
5 of 671 tracked benchmark slots have displayable evidence · bars run 0–100
Coverage detailsHow much of this is verified
Coverage is split by category so a strong number never hides a thin evidence base. Verified means the row is tied to a published source; provisional rows remain visible but separate.
- Agentic3/3 verified
- Coding2/2 verified
- ReasoningNot measured
- MultimodalNot measured
- KnowledgeNot measured
- MultilingualNot measured
- Inst. FollowingNot measured
- MathNot measured
Verified sourceProvisionalNot measured
Benchmark ledger
Coding opens by default. The marker compares each value with the best source-verified result in the catalog; provisional leaders do not set the reference. Expand the remaining categories for every published row.
Coding2 rows
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| DeepSWE | Score61.5% | Versus best verified row Best verified: Muse Wonder Full · 79.0% | Gap17.5 behind | Weight15% ref. weight | Provider exact |
| Terminal-Bench 2.1 | Score83.2% | Versus best verified row Best verified: SWE-2 · 92.8% | Gap9.6 behind | WeightScored in Agentic | Provider exact |
Agentic3 rows
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| Terminal-Bench 2.1 | Score83.2% | Versus best verified row Best verified: SWE-2 · 92.8% | Gap9.6 behind | Weight8% ref. weight | Provider exact |
| JobBench | Score62.9% | Versus best verified row Best verified: Muse Wonder Full · 75.3% | Gap12.4 behind | Weight6% ref. weight | Provider exact |
| AutomationBench | Score50.2% | Versus best verified row Best verified: DeepSeek V4.1 Flash · 54.8% | Gap4.6 behind | Weight6% ref. weight | Provider exact |
Bars run 0–100; the dark tick marks the best source-verified value
All 5 rowsLineage
The sequence follows explicit supersedes links. Each score is estimated for that model; a relative can inform a sparse estimate but never sets a floor, so a newer release can score below an earlier one. Scores and prices remain blank when the corresponding public row or first-party rate is unavailable.
Flash
Radar
Full release historyUltramarine 1 Flash release history
Radar confirmed these at the source. Use Ultramarine 1 Flash in your work? Explore Radar to follow supported changes and choose your alerts.
Radar
Spec sheet
Each documented value carries its source. Missing fields stay visible as not sourced or not published, rather than disappearing from the page.
- API model ID
- ultramarine-1-flashPaper Instruments Ultramarine API
- Context window
- Not publishedPaper Instruments Ultramarine API
- Maximum output
- 32,768 tokens including reasoningPaper Instruments Ultramarine API
- Knowledge cutoff
- Not sourced yet
- Input modalities
- text, imagePaper Instruments Ultramarine API
- Output modalities
- textPaper Instruments Ultramarine API
- Parameters
- Not disclosed by the provider
- Availability
- Paper Instruments APIPaper Instruments Ultramarine availability
- Cloud regions
- Not tracked yet
- Lifecycle
- activePaper Instruments Ultramarine API
- API capabilities
- Chat Completions, Responses, and Messages endpoints; streaming, function tools, and JSON schema output. Reasoning is required at low, high, or max effort.Paper Instruments Ultramarine API
- Prompt caching
- Published at $0.030 per million cached input tokensPaper Instruments public pricing endpoint
- Self-host
- Weights are not published
- Rate limits
- Not tracked yet
How to read this profile
The visual layer above carries the decisions. These notes preserve the model, ranking, coverage, and family context behind the numbers.
We track Ultramarine 1 Flash, but the public leaderboard excludes this profile until enough non-generated benchmark coverage is available. Only published rows appear above.
Ultramarine 1 Flash is a proprietary model. It uses an explicit reasoning mode, which can improve complex problem solving while adding latency and token use.
Available through the Paper Instruments API with text and image inputs, text output, function tools, and structured output. Reasoning is required; low, high, and max effort are supported. The output ceiling is 32,768 tokens including reasoning. The product page says model weights are coming soon; no downloadable checkpoint or license is established.
Paper Instruments post-trained GLM-5.3 Flash for professional knowledge work. Its October 14 report evaluates Ultramarine at maximum reasoning effort. JobBench, AutomationBench, Terminal-Bench 2.1, and DeepSWE 1.1 are recorded with provider attribution and harness details. The report also gives 57.0 Pass@1 on APEX-Agents 1.1 using Mercor Loop over three runs; this stays separate from original APEX-Agents. FrontierFinance 53.3 uses the Samaya Finance Agent v2 R_all metric over 220 queries with a 300-second tool budget, rather than the Vals Finance Agent v2 lane. PI Knowledge Work 59.3 and DiligenceBench 50.6 are publisher evaluations. The 262,144-token window describes its internal Feather evaluation; a hosted API context limit is not established.
5 of 671 tracked benchmark slots currently have displayable evidence. Missing categories stay blank.
Last updated October 14, 2026. Runtime fields remain blank until a sourced snapshot exists.
Questions
How does Ultramarine 1 Flash perform overall in AI benchmarks?
Ultramarine 1 Flash has 5 source-displayable benchmark rows, but it does not qualify for a public overall rank. The available rows remain visible by category without being converted into a site-wide score. Missing evidence stays blank instead of being estimated from an earlier model.
Is Ultramarine 1 Flash good for coding and programming?
Ultramarine 1 Flash has source-displayable benchmark coverage for coding and programming, but the public category table does not assign it a rank there. The individual rows remain available for inspection. A missing category position means the evidence threshold was not met; it does not convert the model's unmeasured work into a zero.
Is Ultramarine 1 Flash good for agentic tool use and computer tasks?
Ultramarine 1 Flash has source-displayable benchmark coverage for agentic tool use and computer tasks, but the public category table does not assign it a rank there. The individual rows remain available for inspection. A missing category position means the evidence threshold was not met; it does not convert the model's unmeasured work into a zero.
Does Ultramarine 1 Flash have full benchmark coverage on BenchLM?
No. Ultramarine 1 Flash currently has 5 source-displayable rows across 671 tracked benchmark slots. The profile exposes published, non-generated evidence and leaves missing categories blank until an exact evaluation is available. Coverage describes how much was measured; it is not a penalty added to an individual benchmark result.
What is the context window size of Ultramarine 1 Flash?
Ultramarine 1 Flash's provider has not published a context-window limit for this exact model. The profile leaves the value unavailable instead of borrowing a number from an earlier family member or a third-party route. Maximum output length is tracked separately when the provider documents it.