Shared benchmarks

27
Shared benchmarks for Kimi K2.5 and Kimi K2.6
BenchmarkKimi K2.5Kimi K2.6
  • 60.6%
  • 83.2%
Evaluation details for BrowseComp

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Kimi K2.5

Record for Kimi K2.5 on BrowseComp BrowseComp

Kimi K2.6

Record for Kimi K2.6 on BrowseComp BrowseComp
  • 12.6%
  • 12.2%
Evaluation details for Chartography October 2026 leaderboard

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Kimi K2.5

Record for Kimi K2.5 on Chartography October 2026 leaderboard

Kimi K2.6

Record for Kimi K2.6 on Chartography October 2026 leaderboard
  • 77.5%
  • 80.4%
Evaluation details for CharXiv Reasoning — No tools

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Kimi K2.5

Record for Kimi K2.5 on CharXiv Reasoning — No tools

Kimi K2.6

Record for Kimi K2.6 on CharXiv Reasoning — No tools
  • 35.8%
  • 44.9%
Evaluation details for Finance Agent Benchmark 2 — October 2026 evaluation

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

  • 87.6%
  • 90.5%
Evaluation details for GPQA Diamond

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Kimi K2.5

Record for Kimi K2.5 on Graduate-Level Google-Proof Q&A Diamond

Kimi K2.6

Record for Kimi K2.6 on Graduate-Level Google-Proof Q&A Diamond
  • 87.6%
  • 90.8%
Evaluation details for GPQA Diamond — Epoch AI harness

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Kimi K2.5

Record for Kimi K2.5 on Graduate-Level Google-Proof Q&A Diamond — Epoch AI harness

Kimi K2.6

Record for Kimi K2.6 on Graduate-Level Google-Proof Q&A Diamond — Epoch AI harness
  • 84.1%
  • 89.1%
Evaluation details for GPQA Diamond — Vals AI harness

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Kimi K2.5

Record for Kimi K2.5 on Graduate-Level Google-Proof Q&A Diamond — Vals AI harness

Kimi K2.6

Record for Kimi K2.6 on Graduate-Level Google-Proof Q&A Diamond — Vals AI harness
  • 0.0%
  • 1.7%
Evaluation details for HLAB October 2026 evaluation

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Kimi K2.5

Record for Kimi K2.5 on Harvey's Legal Agent Benchmark October 2026 evaluation

Kimi K2.6

Record for Kimi K2.6 on Harvey's Legal Agent Benchmark October 2026 evaluation
  • 30.1%
  • 34.7%
Evaluation details for Humanity's Last Exam Full set — No tools

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Kimi K2.5

Record for Kimi K2.5 on Humanity's Last Exam Full set — No tools

Kimi K2.6

Record for Kimi K2.6 on Humanity's Last Exam Full set — No tools
  • 50.2%
  • 54.0%
Evaluation details for Humanity's Last Exam Full set — With tools

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Kimi K2.5

Record for Kimi K2.5 on Humanity's Last Exam Full set — With tools

Kimi K2.6

Record for Kimi K2.6 on Humanity's Last Exam Full set — With tools
  • 85.9%
  • 87.6%
Evaluation details for MMLU-Pro Original — Vals AI harness

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Kimi K2.5

Record for Kimi K2.5 on MMLU-Pro Original — Vals AI harness

Kimi K2.6

Record for Kimi K2.6 on MMLU-Pro Original — Vals AI harness
  • 78.5%
  • 79.4%
Evaluation details for MMMU-Pro No tools

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Kimi K2.5

Record for Kimi K2.5 on MMMU-Pro No tools

Kimi K2.6

Record for Kimi K2.6 on MMMU-Pro No tools
  • 84.3%
  • 86.3%
Evaluation details for MMMU-Pro Vals AI harness

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Kimi K2.5

Record for Kimi K2.5 on MMMU-Pro Vals AI harness

Kimi K2.6

Record for Kimi K2.6 on MMMU-Pro Vals AI harness
OSWorldPotentially non-equivalent: Benchmark versions differ. Evaluator sets differ.
  • 63.3%Verified — 100 steps · XLang Lab
  • 73.1%Verified · Moonshot AI
Potentially non-equivalent · Evaluation details for OSWorld

Benchmark versions differ. Evaluator sets differ. Evaluation methodology is not recorded by the registry.

Kimi K2.5

Record for Kimi K2.5 on OSWorld Verified — 100 steps

Kimi K2.6

Record for Kimi K2.6 on OSWorld Verified
  • 12.0%
  • 7.2%
Evaluation details for Riemann-bench October 2026 leaderboard

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Kimi K2.5

Record for Kimi K2.5 on Riemann-bench October 2026 leaderboard

Kimi K2.6

Record for Kimi K2.6 on Riemann-bench October 2026 leaderboard
  • 34.3%
  • 34.9%
Evaluation details for SimpleQA Verified Original — Epoch AI harness

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

  • 73.0%
  • 76.7%
Evaluation details for SWE-bench Multilingual

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Kimi K2.5

Record for Kimi K2.5 on SWE-bench Multilingual

Kimi K2.6

Record for Kimi K2.6 on SWE-bench Multilingual
  • 76.8%
  • 80.2%
Evaluation details for SWE-bench Verified

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Kimi K2.5

Record for Kimi K2.5 on SWE-bench Verified

Kimi K2.6

Record for Kimi K2.6 on SWE-bench Verified
  • 73.8%
  • 76.7%
Evaluation details for SWE-bench Verified — Epoch AI harness

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Kimi K2.5

Record for Kimi K2.5 on SWE-bench Verified — Epoch AI harness

Kimi K2.6

Record for Kimi K2.6 on SWE-bench Verified — Epoch AI harness
  • 70.0%
  • 76.2%
Evaluation details for SWE-bench Verified — Vals AI mini-swe-agent

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Kimi K2.5

Record for Kimi K2.5 on SWE-bench Verified — Vals AI mini-swe-agent

Kimi K2.6

Record for Kimi K2.6 on SWE-bench Verified — Vals AI mini-swe-agent
  • 50.7%
  • 58.6%
Evaluation details for SWE-bench Pro Public

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Kimi K2.5

Record for Kimi K2.5 on SWE-bench Pro Public

Kimi K2.6

Record for Kimi K2.6 on SWE-bench Pro Public
  • 50.8%
  • 66.7%
Evaluation details for Terminal-Bench 2.0

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Kimi K2.5

Record for Kimi K2.5 on Terminal-Bench 2.0

Kimi K2.6

Record for Kimi K2.6 on Terminal-Bench 2.0
  • 40.4%
  • 57.3%
Evaluation details for Terminal-Bench 2.0 — Vals AI Terminus 2

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Kimi K2.5

Record for Kimi K2.5 on Terminal-Bench 2.0 — Vals AI Terminus 2

Kimi K2.6

Record for Kimi K2.6 on Terminal-Bench 2.0 — Vals AI Terminus 2
  • 41.9%
  • 53.6%
Evaluation details for Terminal-Bench 2.1 — Vals AI Terminus 2

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Kimi K2.5

Record for Kimi K2.5 on Terminal-Bench 2.1 — Vals AI Terminus 2

Kimi K2.6

Record for Kimi K2.6 on Terminal-Bench 2.1 — Vals AI Terminus 2
Toolathlon OriginalPotentially non-equivalent: Evaluator sets differ.
  • 27.8%Original · Toolathlon Authors
  • 50.0%Original · Moonshot AI
Potentially non-equivalent · Evaluation details for Toolathlon Original

Evaluator sets differ. Evaluation methodology is not recorded by the registry.

Kimi K2.5

Record for Kimi K2.5 on Toolathlon Original

Kimi K2.6

Record for Kimi K2.6 on Toolathlon Original
  • 33.0%
  • 58.0%
Evaluation details for Toolathlon Verified

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Kimi K2.5

Record for Kimi K2.5 on Toolathlon Verified

Kimi K2.6

Record for Kimi K2.6 on Toolathlon Verified
  • 17.5%
  • 37.9%
Evaluation details for Vibe Code Bench 1.1 — OpenHands, October 2026 evaluation

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.