Language modelActive
Nex-N2-mini
Nex AGI
- Released
- June 4, 2026
- Data date
- September 8, 2026
- Predecessor
- Nex-N1 (Official documentation)
- Price
- Historical rates under Costs
- Input
- Text
- Output
- Text
Nex-N2-mini combines 35 billion total parameters with about 3 billion active and Nex AGI's agentic thinking approach. Built on Qwen 3.5 35B A3B, its Adaptive Thinking stays brief on simple actions and works more deeply on critical decisions. Coherent Thinking is intended to keep that shift consistent across general and agentic tasks.
Nex AGI presents Mini alongside Nex-N2-Pro as a pair with different model sizes and compute requirements. Mini activates about 3 billion parameters per token, while Pro activates about 17 billion.
Specifications and access
| Specification | Value and source |
|---|---|
| Weights | Open weightsSource |
| Access | LocalSource |
| Parameters | 35BSource |
| Active parameters | 3BSource |
| Context window | 262,144 tokensSource |
| License | Apache-2.0Source |
| Parameters | 35 BSource |
| Active parameters | 3 BSource |
| Architecture | Qwen3.5-35B-A3B-Base post-trained with Nex Agentic ThinkingSource |
| Input | TextSource |
| Output | TextSource |
| Reasoning mode | Adaptive Thinking and Coherent Thinking; model decides when and how deeply to thinkSource |
| Tool calling | YesSource |
Page 1 of 2
Measurements without matching peer values
These measurements have no matching peer values under the same test conditions. Their original values and sources remain available here.
BrowseComp (Show measurement, test conditions, and source)
- Source value
- 74.1
- Score
- 74.1
- Metric
- percent
- Unit
- %
- Category
- agentic
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Nex-N2 model card
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Vendor-reported result. Vendor comparison table; exact shots and harness configuration are not stated in the visible table.
Toolathlon (Show measurement, test conditions, and source)
- Source value
- 33.3
- Score
- 33.3
- Metric
- percent
- Unit
- %
- Category
- agentic
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Nex-N2 model card
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Vendor-reported result. Vendor comparison table; exact harness configuration not stated.
WildClawBench (Show measurement, test conditions, and source)
- Source value
- 47.7
- Score
- 47.7
- Metric
- percent
- Unit
- %
- Category
- agentic
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Nex-N2 model card
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Vendor-reported result. Vendor comparison table; exact harness configuration not stated.
WideSearch (Show measurement, test conditions, and source)
- Source value
- 62
- Score
- 62
- Metric
- percent
- Unit
- %
- Category
- agentic
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Nex-N2 model card
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Vendor-reported result. Vendor comparison table; exact harness configuration not stated.
TAU3 (Show measurement, test conditions, and source)
- Source value
- 65.9
- Score
- 65.9
- Metric
- percent
- Unit
- %
- Category
- agentic
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Nex-N2 model card
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Vendor-reported result. Vendor comparison table; exact harness configuration not stated.
SWE-Bench Pro (Show measurement, test conditions, and source)
- Source value
- 50.2
- Score
- 50.2
- Metric
- percent
- Unit
- %
- Category
- coding
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Nex-N2 model card
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Vendor-reported result. Vendor comparison table; exact harness configuration not stated.
Terminal-Bench 2.1 (Show measurement, test conditions, and source)
- Source value
- 60.7
- Score
- 60.7
- Metric
- percent
- Unit
- %
- Category
- coding
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Nex-N2 model card
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Vendor-reported result. Vendor comparison table; exact harness configuration not stated.
DeepSWE (Show measurement, test conditions, and source)
- Source value
- 8
- Score
- 8
- Metric
- percent
- Unit
- %
- Category
- coding
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Nex-N2 model card
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Vendor-reported result. Vendor comparison table; exact harness configuration not stated.
SWE Verified (Show measurement, test conditions, and source)
- Source value
- 74.4
- Score
- 74.4
- Metric
- percent
- Unit
- %
- Category
- coding
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Nex-N2 model card
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Vendor-reported result. Vendor comparison table; exact harness configuration not stated.
SWE Atlas QnA (Show measurement, test conditions, and source)
- Source value
- 31.5
- Score
- 31.5
- Metric
- percent
- Unit
- %
- Category
- coding
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Nex-N2 model card
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Vendor-reported result. Vendor comparison table; exact harness configuration not stated.
SWE Atlas RF (Show measurement, test conditions, and source)
- Source value
- 30
- Score
- 30
- Metric
- percent
- Unit
- %
- Category
- coding
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Nex-N2 model card
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Vendor-reported result. Vendor comparison table; exact harness configuration not stated.
SWE Atlas TW (Show measurement, test conditions, and source)
- Source value
- 23.3
- Score
- 23.3
- Metric
- percent
- Unit
- %
- Category
- coding
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Nex-N2 model card
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Vendor-reported result. Vendor comparison table; exact harness configuration not stated.
Page 1 of 2
Popularity
34.7 / 100
Predecessor and successor
Sources and data date
| Type | Evidence and data date |
|---|---|
| Input | Documented modalities. Retrieved September 8, 2026. |
| Output | Documented modalities. Retrieved September 8, 2026. |
| Research status | Research date September 8, 2026. 3 source URLs checked. This documents the inspected sources, not an exhaustive inventory of every publication. |
| Not found in the inspected sources | knowledge-cutoff Original wording The exact Nex-N2-mini card does not publish a knowledge cutoff. |
| Unresolved | AA benchmark rows Original wording No exact Artificial Analysis model page located in the checked public sitemap. |
| Additional source | API pricing |
| Additional source | OpenRouter: Popularity index (derived), CC BY 4.0 |
| Additional source | Vercel AI Gateway: Popularity index (derived), CC BY 4.0 |
| Additional source | Creative Commons Attribution 4.0 |
| Additional source | Nex-N2-Pro model card (Hugging Face) (2026-09-19) |