Language modelOpen weights

Xing4.0-29B-A4B

XingChen AGI

Released: -
Data date: October 3, 2026

Xing4.0-29B-A4B was trained entirely on Ascend NPUs using MindSpore. Its developer adapted the MoE architecture to Ascend 910C clusters for this purpose. The model activates 4 of its 29 billion parameters per token.

Its standard context window holds 256K tokens and can be extended to 512K. The weights use Apache 2.0, and the model card documents local runtimes and an OpenAI-compatible API.

Specifications and access

Specification	Value and source
Model ID	XingChen-AGI/Xing4.0-29B-A4BSource
Parameters	29 billion total, 4 billion activeSource
Context window	256K tokens, extensible to 512KSource
Architecture	mHC, MLA, and MTPSource
Access	OpenAI-compatible API and local deploymentSource
License	Apache 2.0Source

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.

IFBench (Show measurement, test conditions, and source)

Source value: 69.67
Score: 69.67
Metric: IFBench reported score
Unit: No unit provided
Category: source-specific
Direction: Higher is better
Source type: vendor-reported
Evaluator: Provider model card
Status: active
Retrieved at: 2026-10-04
Methodology: Vendor-reported result. Vendor card does not publish sufficient setup/sampling detail; many rows are agent/harness tasks.
Context: Source-specific observation; it is not a shared comparison cohort. The source table does not state a unit or scale for this numeric score.

Source Provider model card

AIME 2026 (Show measurement, test conditions, and source)

Source value: 90
Score: 90
Metric: AIME 2026 reported score
Unit: No unit provided
Category: source-specific
Direction: Higher is better
Source type: vendor-reported
Evaluator: Provider model card
Status: active
Retrieved at: 2026-10-04
Methodology: Vendor-reported result. Vendor card does not publish sufficient setup/sampling detail; many rows are agent/harness tasks.
Context: Source-specific observation; it is not a shared comparison cohort. The source table does not state a unit or scale for this numeric score.

Source Provider model card

AA.LCR (Show measurement, test conditions, and source)

Source value: 61
Score: 61
Metric: AA.LCR reported score
Unit: No unit provided
Category: source-specific
Direction: Higher is better
Source type: vendor-reported
Evaluator: Provider model card
Status: active
Retrieved at: 2026-10-04
Methodology: Vendor-reported result. Vendor card does not publish sufficient setup/sampling detail; many rows are agent/harness tasks.
Context: Source-specific observation; it is not a shared comparison cohort. The source table does not state a unit or scale for this numeric score.

Source Provider model card

tau3Bench (Show measurement, test conditions, and source)

Source value: 64.63
Score: 64.63
Metric: tau3Bench reported score
Unit: No unit provided
Category: source-specific
Direction: Higher is better
Source type: vendor-reported
Evaluator: Provider model card
Status: active
Retrieved at: 2026-10-04
Methodology: Vendor-reported result. Vendor card does not publish sufficient setup/sampling detail; many rows are agent/harness tasks.
Context: Source-specific observation; it is not a shared comparison cohort. The source table does not state a unit or scale for this numeric score.

Source Provider model card

ClawEval (Show measurement, test conditions, and source)

Source value: 76.55
Score: 76.55
Metric: ClawEval reported score
Unit: No unit provided
Category: source-specific
Direction: Higher is better
Source type: vendor-reported
Evaluator: Provider model card
Status: active
Retrieved at: 2026-10-04
Methodology: Vendor-reported result. Vendor card does not publish sufficient setup/sampling detail; many rows are agent/harness tasks.
Context: Source-specific observation; it is not a shared comparison cohort. The source table does not state a unit or scale for this numeric score.

Source Provider model card

SWE-Verified (Show measurement, test conditions, and source)

Source value: 75
Score: 75
Metric: SWE-Verified reported score
Unit: No unit provided
Category: source-specific
Direction: Higher is better
Source type: vendor-reported
Evaluator: Provider model card
Status: active
Retrieved at: 2026-10-04
Methodology: Vendor-reported result. Vendor card does not publish sufficient setup/sampling detail; many rows are agent/harness tasks.
Context: Source-specific observation; it is not a shared comparison cohort. The source table does not state a unit or scale for this numeric score.

Source Provider model card

TerminalBench 2.1 (Show measurement, test conditions, and source)

Source value: 57.5
Score: 57.5
Metric: TerminalBench 2.1 reported score
Unit: No unit provided
Category: source-specific
Direction: Higher is better
Source type: vendor-reported
Evaluator: Provider model card
Status: active
Retrieved at: 2026-10-04
Methodology: Vendor-reported result. Vendor card does not publish sufficient setup/sampling detail; many rows are agent/harness tasks.
Context: Source-specific observation; it is not a shared comparison cohort. The source table does not state a unit or scale for this numeric score.

Source Provider model card

SWE Multilingual (Show measurement, test conditions, and source)

Source value: 66
Score: 66
Metric: SWE Multilingual reported score
Unit: No unit provided
Category: source-specific
Direction: Higher is better
Source type: vendor-reported
Evaluator: Provider model card
Status: active
Retrieved at: 2026-10-04
Methodology: Vendor-reported result. Vendor card does not publish sufficient setup/sampling detail; many rows are agent/harness tasks.
Context: Source-specific observation; it is not a shared comparison cohort. The source table does not state a unit or scale for this numeric score.

Source Provider model card

DeepResearchBII (Show measurement, test conditions, and source)

Source value: 60.8
Score: 60.8
Metric: DeepResearchBII reported score
Unit: No unit provided
Category: source-specific
Direction: Higher is better
Source type: vendor-reported
Evaluator: Provider model card
Status: active
Retrieved at: 2026-10-04
Methodology: Vendor-reported result. Vendor card does not publish sufficient setup/sampling detail; many rows are agent/harness tasks.
Context: Source-specific observation; it is not a shared comparison cohort. The source table does not state a unit or scale for this numeric score.

Source Provider model card

Sources and data date

Every statement links to its underlying documentation or leaderboard.

Type	Evidence and data date
Research status	Research date October 4, 2026. 1 source URLs checked. This documents the inspected sources, not an exhaustive inventory of every publication.
Additional source	Xing4.0-29B-A4B model card (retrieved October 3, 2026; October 4, 2026) · Xing4.0-29B-A4B model card · Editorial description reviewed October 4, 2026