Data as of October 29, 2026 · How the score is built
Released Sep 14, 2026256K contextAtria Dawn Preview model card
Atria Dawn Preview
Decision readingAtria Dawn Preview is tracked, but not publicly ranked yet. The profile exposes 13 sourced benchmark rows and leaves unsupported fields blank until a published record exists.
Released Sep 14, 2026 — see all recent releases
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 50.5Not eligible for a public rank
Price
Self-hosted; infrastructure cost varies
input median $0.75No comparable first-party hosted token rate
Speed
Not measured
field median 92 tok/sTime to first token not measured
Context
256Ktokens
field median 256,000Maximum output length is tracked separately
Strongest published evidence
Published rows are visible, but no category has enough eligible evidence for a comparative rank.
Validate before choosing
13 published rows leave some tracked benchmark slots empty. No comparable first-party API token rate is published.
Source-linked · 13 displayable benchmark rows
Category 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%11 benchmarksVerified | 82.3 | 11 benchmarks | Verified | |||
| CodingRank Not rankedWeight 20%2 benchmarksVerified | 48.7 | 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 |
13 of 782 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.
- Agentic11/11 verified
- Coding2/2 verified
- ReasoningNot measured
- MultimodalNot measured
- KnowledgeNot measured
- MultilingualNot measured
- Inst. FollowingNot measured
- MathNot 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 |
|---|---|---|---|---|---|
| SWE-bench Pro | Score59.6% | Versus best verified row Best verified: Claude Opus 5.5 · 89.9% | Gap30.3 behind | Weight26% ref. weight | Provider exact |
| Terminal-Bench 2.1 | Score78.3% | Versus best verified row Best verified: SWE-2 · 92.8% | Gap14.5 behind | WeightScored in Agentic | Provider exact |
Agentic11 rows
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| BrowseComp | Score92.5% | Versus best verified row Best verified: Atria Dawn Preview · 92.5% | GapBest verified | Weight8% ref. weight | Provider exact |
| Terminal-Bench 2.1 | Score78.3% | Versus best verified row Best verified: SWE-2 · 92.8% | Gap14.5 behind | Weight8% ref. weight | Provider exact |
| AutomationBench | Score53.8% | Versus best verified row Best verified: MiniMax M3.1 Flash · 62.2% | Gap8.4 behind | Weight6% ref. weight | Provider exact |
| JobBench | Score50.3% | Versus best verified row Best verified: Muse Wonder Full · 75.3% | Gap25 behind | Weight6% ref. weight | Provider exact |
| BFCL v4Berkeley Function Calling Leaderboard v4 | Score77.0% | Versus best verified row Best verified: BTL-3 · 88.5% | Gap11.5 behind | Weight3% ref. weight | Provider exact |
| DeepSearchQA | Score96.0% | Versus best verified row Best verified: Atria Dawn Preview · 96.0% | GapBest verified | Weight2% ref. weight | Provider exact |
| τ³-bench resultsτ³-Bench Tool-Agent-User Evaluation | Score41.2% | Versus best verified row Best verified: Mercury 2.5 · 96.0% | Gap54.8 behind | Weight2% ref. weight | Provider exact |
| CyberGym | Score86.5% | Versus best verified row Best verified: MiMo-V2.6-Flash · 95.1% | Gap8.6 behind | WeightDisplay only | Provider exact |
| skillsBench | Score66.4% | Versus best verified row Best verified: Qwen3.8 Max · 70.2% | Gap3.8 behind | WeightDisplay only | Provider exact |
| MLE-Bench Lite | Score86.2% | Versus best verified row Best verified: Atria Dawn Preview · 86.2% | GapBest verified | WeightDisplay only | Provider exact |
| WideResearch | Score81.9% | Versus best verified row Best verified: Hy4 preview · 83.9% | Gap2 behind | WeightDisplay only | Provider exact |
Bars run 0–100; the dark tick marks the best source-verified value
All 13 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.
preview · Preview
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
- Atria-Dawn-PreviewAtria Dawn Preview model card
- Context window
- 256KAtria Dawn Preview model card
- Maximum output
- Not sourced yet
- Knowledge cutoff
- Not sourced yet
- Input modalities
- Not sourced yet
- Output modalities
- Not sourced yet
- Parameters
- Not sourced yet
- Availability
- Shanghai Artificial Intelligence Laboratory publishes the full-precision and FP8 Atria Dawn Preview checkpoints under MIT for self-hosting. Atria also provides hosted access through Chat Completions, Messages, and Responses-compatible APIs with the exact model ID `Atria-Dawn-Preview`; no first-party token rate is published.
- Cloud regions
- Not tracked yet
- Lifecycle
- Current
- API capabilities
- Tool calling, structured outputs, and batch support are not tracked yet
- Prompt caching
- Not documented in the pricing recordAtria Dawn Preview model card
- Self-host
- Atria Dawn Preview needs ~640GB VRAM (8× NVIDIA H100 (80GB)).
- 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 Atria Dawn Preview, but the public leaderboard excludes this profile until enough non-generated benchmark coverage is available. Only published rows appear above.
Atria Dawn Preview is a open weight model with a 256K context window. It uses an explicit reasoning mode, which can improve complex problem solving while adding latency and token use.
Shanghai Artificial Intelligence Laboratory publishes the full-precision and FP8 Atria Dawn Preview checkpoints under MIT for self-hosting. Atria also provides hosted access through Chat Completions, Messages, and Responses-compatible APIs with the exact model ID `Atria-Dawn-Preview`; no first-party token rate is published.
13 of 782 tracked benchmark slots currently have displayable evidence. Missing categories stay blank.
Last updated October 29, 2026. Runtime fields remain blank until a sourced snapshot exists.
Deployment options
Self-host and provider-specific paths stay separate from benchmark evidence so operating constraints are visible before a score becomes the whole decision.
Self-host vs API cost
Estimates at 50,000 req/day · 1000 tokens/req average.
Questions
How does Atria Dawn Preview perform overall in AI benchmarks?
Atria Dawn Preview has 13 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 Atria Dawn Preview good for coding and programming?
Atria Dawn Preview 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 Atria Dawn Preview good for agentic tool use and computer tasks?
Atria Dawn Preview 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.
Is Atria Dawn Preview open source?
Atria Dawn Preview is an open-weight model from Shanghai Artificial Intelligence Laboratory. Its weights can be downloaded for local or hosted deployment, subject to the published license. Open weight does not automatically mean open source: training data and training code may remain private, and commercial restrictions can still apply.
Does Atria Dawn Preview have full benchmark coverage on BenchLM?
No. Atria Dawn Preview currently has 14 source-displayable rows. 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 Atria Dawn Preview?
Atria Dawn Preview has a documented context window of 256K. That figure is the maximum combined prompt and retained-conversation space reported for this exact model; it is not the maximum output length. The profile keeps output limits separate because providers often publish those limits independently.