readable AI benchmarks (simplified)

readable AI benchmarks

MBZUAI Institute of Foundation Models

K2 Horizon 375B A23B

reasoning · open weights · Sep 3, 2026

K2 Horizon 375B A23B has a Quality Score of 47.1, ranking 65th among 278 scored models. It accepts text input, outputs text, and has a 524,288 token context window. Its Reliability Score is 48.5, ranking 64th among 278 scored models, above average (total average: 34.7).

Quality Score47.1coverage 91.2%: missing Epoch General ECI, CursorBench 3.2 and DeepSWE v1.1
Value Scorenot available
Reliability48.5
Cache Discountnot available

Model specification

Reasoning
reasoning
Input modalities
text
Output modalities
text
Context window
524,288 tokens
Weights
open weights

Token prices USD per 1M tokens

Input
not available
Output
not available
Cached Input
not available
Cached Output
not available
Cost per task
not available

Capability

Intelligence
47.3
Coding
61.5
Agentic
42.8
Omniscience
-3.0
Correct
18.2%
Blended price
not available

Answer outcomes

Attempt rate: 39.3%

Correct 18.2%Incorrect 21.2%Abstained 60.7%

Artificial Analysis benchmarks

GDPval-AA v2
46%
τ³-Banking
34%
Terminal-Bench v2.1
72%
SciCode
41%
Humanity’s Last Exam
32%
GPQA Diamond
87%
CritPt
5%
AA-Omniscience
49%
AA-LCR
76%

Similar models