MBZUAI Institute of Foundation ModelsOpen weightsReasoning modelAuto-detected

K2 Think V2

Name: K2 Think V2 benchmark results
Creator: BenchLeader

Best configuration ranks #309 of 610 on the BenchLeader Index at 49.3 ±7.2 (thinking reasoning effort).

Blended price: –
Output speed: –
First answer: –
Context: 262k

How it scores by categoryDashed line = average model (50). One step of 15 = one standard deviation.

Overall index49
Agents & tools40
Knowledge45
Instruction following62
Long context55
Composite44

Reasoning-effort configurations

The same model behaves differently depending on how much it is allowed to think. Each row is one setting, scored only on the benchmarks that were run at that setting. “Default” means the publisher did not say which setting was used.

Effort	Index	Rank	Speed	First answer	Chat reply cost	Categories
low	43.2	#441	–	–	–	Agents & tools 38 · Composite 38 · Instruction following 43 · Knowledge 42 · Long context 35
medium	45.8	#383	–	–	–	Agents & tools 41 · Composite 40 · Instruction following 55 · Knowledge 41 · Long context 39
thinkingbest	49.3	#309	–	–	–	Agents & tools 40 · Composite 44 · Instruction following 62 · Knowledge 45 · Long context 55
default	46.7	#364	–	–	–	Agents & tools 42 · Composite 41 · Instruction following 60 · Knowledge 38 · Long context 43

Benchmark results

One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.

Reasoning

Benchmark	low	medium	thinking	default	Source
GPQA Diamond (AA)not in index	54.1%#388	59.8%#359	71.3%#277	68.1%#302	Artificial Analysis
Humanity's Last Exam (AA)not in index	3.6%#497	4.4%#427	10.1%#258	10.5%#249	Artificial Analysis

Agents & tools

Benchmark	low	medium	thinking	default	Source
Terminal-Bench Hard	4.5%#270	8.3%#226	6.8%#240	9.8%#220	Artificial Analysis
τ²-Bench Telecom (AA)not in index	20.8%#315	24.9%#285	25.4%#280	27.8%#264	Artificial Analysis

Knowledge

Benchmark	low	medium	thinking	default	Source
AA-Omniscience	-47.3#294	-49#309	-40.3#255	-55.6#363	Artificial Analysis

Instruction following

Benchmark	low	medium	thinking	default	Source
IFBench	41.0%#246	55.1%#145	62.8%#115	60.1%#124	Artificial Analysis

Long context

Benchmark	low	medium	thinking	default	Source
AA-LCR	20.3%#360	28.0%#328	57.0%#230	35.0%#303	Artificial Analysis

Composite

Benchmark	low	medium	thinking	default	Source
AA Intelligence Index	7.3#403	9.0#339	11.5#284	9.9#317	Artificial Analysis