RLTasksby ulam.ai
Models / Kimi

Kimi / MODEL EVIDENCE

Kimi K2

4 benchmarks · 12 observations · 4 sources · released 2025-07-11

Scores at a glance

4 benchmarks · large dot: Kimi K2; small dots: closest models where reported

GAIA-2 ↗34.2%
7.169.2

Kimi-K2 · GAIA-2 · Gaia2 full · Execution pass@1 · baseline ReAct · three runs ↗

7 more reported results

36% · Kimi-K2 · GAIA-2 · Gaia2 full · Search pass@1 · baseline ReAct · three runs · Gaia2: Benchmarking LLM Agents on Dynamic and Asynchronous Environments ↗

8.3% · Kimi-K2 · GAIA-2 · Gaia2 full · Ambiguity pass@1 · baseline ReAct · three runs · Gaia2: Benchmarking LLM Agents on Dynamic and Asynchronous Environments ↗

24% · Kimi-K2 · GAIA-2 · Gaia2 full · Adaptability pass@1 · baseline ReAct · three runs · Gaia2: Benchmarking LLM Agents on Dynamic and Asynchronous Environments ↗

0.8% · Kimi-K2 · GAIA-2 · Gaia2 full · Time pass@1 · baseline ReAct · three runs · Gaia2: Benchmarking LLM Agents on Dynamic and Asynchronous Environments ↗

18.8% · Kimi-K2 · GAIA-2 · Gaia2 full · Noise pass@1 · baseline ReAct · three runs · Gaia2: Benchmarking LLM Agents on Dynamic and Asynchronous Environments ↗

18.3% · Kimi-K2 · GAIA-2 · Gaia2 full · Agent2Agent pass@1 · baseline ReAct · three runs · Gaia2: Benchmarking LLM Agents on Dynamic and Asynchronous Environments ↗

20.1% · Kimi-K2 · GAIA-2 · Gaia2 full · Overall pass@1 · baseline ReAct · three runs · Gaia2: Benchmarking LLM Agents on Dynamic and Asynchronous Environments ↗

Closest models by score

Similar positions on shared benchmark leaderboards. Smaller gaps are closer.

34.5 percentile-point gap

3 of 4 shared benchmark comparisons

GPT-5 →OpenAI
59.2 percentile-point gap

4 of 4 shared benchmark comparisons

How similarity is measured

Mean absolute gap in percentile positions on shared benchmark settings; lower is closer. At least 60% overlap is required, with a minimum of three when available. Independent evaluations are used when there are at least three; otherwise published report comparisons are used. Repeated model observations within one setting are averaged before ranking. Writing-style traits and composite indices are excluded. This describes similarity, not an overall quality ranking.

Filter comparisons to this model →

Sources

Every plotted observation links to its evaluation. Reported configurations, benchmark versions and task subsets can differ; similarity describes shared scores, not training lift.

  1. Aider code editing benchmark ↗

    Paul Gauthier, Aider contributors

  2. MATH-500 Benchmark | Artificial Analysis ↗

    Artificial Analysis

  3. TextQuests: How Good are LLMs at Text-Based Video Games? — published result table ↗

    Phan, Long, Mazeika, Mantas, Zou, Andy, Hendrycks, Dan · 2025

  4. Gaia2: Benchmarking LLM Agents on Dynamic and Asynchronous Environments — published result table ↗

    Froger, Romain, Andrews, Pierre, Bettini, Matteo, Budhiraja, Amar, Cabral, Ricardo Silveira, Do, Virginie et al. · 2026

    All 24 authors

    Froger, Romain, Andrews, Pierre, Bettini, Matteo, Budhiraja, Amar, Cabral, Ricardo Silveira, Do, Virginie, Garreau, Emilien, Gaya, Jean-Baptiste, Laurençon, Hugo, Lecanu, Maxime, Malkan, Kunal, Mekala, Dheeraj, Ménard, Pierre, Bertran, Gerard Moreno-Torres, Piterbarg, Ulyana, Plekhanov, Mikhail, Rita, Mathieu, Rusakov, Andrey, Vorotilov, Vladislav, Wang, Mengjue, Yu, Ian, Benhalloum, Amine, Mialon, Grégoire, Scialom, Thomas

Names used in source tables

Kimi K2 · Kimi-K2