vs
2:0
overall capability wins (0 ties across 2 scored). Atria Dawn Preview leads overall.
- Specs2:0
Verdict
Atria Dawn Preview offers ~3.9x larger context (256K vs 66K).
Cost for 1M in + 250K out
Illustrative chat workload at primary-provider list prices.
- Atria Dawn Preview
- -
- Olmo 3.1 32B Instruct
- -
Price per 1M tokens
Full comparison
Organization
Shanghai AI Lab
Ai2
Family
Other
Other
License
MIT
Apache 2.0
Open weights
Yes
Yes
Release
Sep 11, 2026
Dec 11, 2025
Knowledge cutoff
-
-
API / provider
No primary API price listed
No primary API price listed
Modalities
text → text
text → text
Specs
Context window
256K
66K
Max output
66K
-
Parameters
744B
32B
Pricing
Input $/1M
—
—
Output $/1M
—
—
Blended $/1M (3∶1)
-
-
Speed
tok/s
-
-
TTFT (s)
-
-
Reasoning
MMLU-Pro
—
—
GPQA Diamond
—
53.9
Humanity's Last Exam
—
5.0
AIME 2025
—
—
MATH-500
—
—
Humanity's Last Exam (with tools)
—
—
AA-Omniscience Accuracy
—
12.0
AA-LCR v1.1
—
—
CritPt
—
—
MMMU-Pro
—
—
IFBench
—
39.2
Chartography
—
—
Chartography (With Tools)
—
—
Coding
SWE-bench Verified
—
—
SWE-bench Pro
59.6
—
SWE-bench Multilingual
—
—
LiveCodeBench
—
—
Terminal-Bench 2.1
—
—
Aider Polyglot
—
—
Terminal-Bench 3
—
—
BigCodeBench
—
—
SciCode
—
—
CursorBench
—
—
SWE-Rebench
—
—
NL2Repo-Bench
—
—
DeepSWE
—
—
WebDev Arena
—
—
Terminal-Bench 4.0
—
—
CursorBench 4.0
—
—
FrontierCode v1.1 (Main)
—
—
Arena
LMArena Elo
—
—
| Metric | Atria Dawn Preview | Olmo 3.1 32B Instruct | Delta |
|---|---|---|---|
| Identity | |||
| Organization | Shanghai AI Lab | Ai2 | — |
| Family | Other | Other | — |
| License | MIT | Apache 2.0 | — |
| Open weights | Yes | Yes | — |
| Release | Sep 11, 2026 | Dec 11, 2025 | — |
| Knowledge cutoff | - | - | — |
| API / provider | No primary API price listed | No primary API price listed | — |
| Modalities | text → text | text → text | — |
| Specs | |||
| Context window | 256K | 66K | A +190K |
| Max output | 66K | - | — |
| Parameters | 744B | 32B | A +712 |
| Pricing | |||
| Input $/1M | — | — | — |
| Output $/1M | — | — | — |
| Blended $/1M (3∶1) | - | - | — |
| Speed | |||
| tok/s | - | - | — |
| TTFT (s) | - | - | — |
| Reasoning | |||
| MMLU-Pro | — | — | — |
| GPQA Diamond | — | 53.9 | — |
| Humanity's Last Exam | — | 5.0 | — |
| AIME 2025 | — | — | — |
| MATH-500 | — | — | — |
| Humanity's Last Exam (with tools) | — | — | — |
| AA-Omniscience Accuracy | — | 12.0 | — |
| AA-LCR v1.1 | — | — | — |
| CritPt | — | — | — |
| MMMU-Pro | — | — | — |
| IFBench | — | 39.2 | — |
| Chartography | — | — | — |
| Chartography (With Tools) | — | — | — |
| Coding | |||
| SWE-bench Verified | — | — | — |
| SWE-bench Pro | 59.6 | — | — |
| SWE-bench Multilingual | — | — | — |
| LiveCodeBench | — | — | — |
| Terminal-Bench 2.1 | — | — | — |
| Aider Polyglot | — | — | — |
| Terminal-Bench 3 | — | — | — |
| BigCodeBench | — | — | — |
| SciCode | — | — | — |
| CursorBench | — | — | — |
| SWE-Rebench | — | — | — |
| NL2Repo-Bench | — | — | — |
| DeepSWE | — | — | — |
| WebDev Arena | — | — | — |
| Terminal-Bench 4.0 | — | — | — |
| CursorBench 4.0 | — | — | — |
| FrontierCode v1.1 (Main) | — | — | — |
| Arena | |||
| LMArena Elo | — | — | — |
Atria Dawn Preview
- Context
- 256K / 66K out
- Parameters
- 744B
- Price
- —
- Speed
- —
- Modalities
- text → text
- License
- MIT
Shanghai AI Laboratory's open-weight agentic MoE (744B total, MIT license, 256K context), built on a GLM-5.2 base and trained by the cross-institutional ATRIA initiative (with Fudan NLP Lab) for long-horizon research and engineering workflows: problem analysis, tool use, code, experiments, and failure recovery. Benchmark scores are from the model's own technical report and have not yet been independently verified by a third-party evaluator.
Olmo 3.1 32B Instruct
- Context
- 66K
- Parameters
- 32B
- Price
- —
- Speed
- —
- Modalities
- text → text
- License
- Apache 2.0
Ai2's strongest fully open 32B-scale instruct model, applying the Olmo 3 Instruct post-training recipe to a larger base; no hosted API yet.
Benchmark charts
Winner bars are emphasized. Per-benchmark deltas sit above each chart.
These two models have no overlapping published benchmarks in our dataset. Compare specs and pricing instead.