Moonshot AI · open weights on Arena · first on Arena 29 January 2026

kimi-k2.5-thinking

219 days for open weights to match gemini-2.5-pro

This open-weights model matched the score of gemini-2.5-pro in Vision 219 days after gemini-2.5-pro took first place.

Embed

Who held first place

<iframe src="https://jesseahawkins.com/frontier/embed/" title="Frontier: who held first place on each Arena leaderboard, and for how long" width="100%" height="560" style="border:0;max-width:100%" loading="lazy"></iframe>

Credits LMArena.

The open-weights catch-up

<iframe src="https://jesseahawkins.com/frontier/embed/catch-up/" title="Frontier: the days until open weights matched each closed leader" width="100%" height="900" style="border:0;max-width:100%" loading="lazy"></iframe>

Credits LMArena.

The saturation clock

<iframe src="https://jesseahawkins.com/frontier/embed/clock/" title="Frontier: the days from each benchmark’s release to a model at 90%" width="100%" height="640" style="border:0;max-width:100%" loading="lazy"></iframe>

Credits Epoch AI.

Each frame links back to its page. Preview it.

Cite

This page

Jesse Hawkins, “kimi-k2.5-thinking · Frontier”, jesseahawkins.com, data retrieved 30 Oct 2026. https://jesseahawkins.com/frontier/kimi-k2-5-thinking/. Data: LMArena, Arena leaderboard dataset, revision 4491b06, CC BY 4.0; Epoch AI, Capabilities & benchmarking, CC BY 4.0.

BibTeX

@misc{hawkins2026frontierkimik25thinking,
  author = {Hawkins, Jesse},
  title = {{kimi-k2.5-thinking · Frontier}},
  year = {2026},
  howpublished = {\url{https://jesseahawkins.com/frontier/kimi-k2-5-thinking/}},
  note = {Data: LMArena, Arena leaderboard dataset, revision 4491b06, CC BY 4.0; Epoch AI, Capabilities \& benchmarking, CC BY 4.0, retrieved 2026-10-30}
}

Arena · LMArena

LMArena, "Arena leaderboard dataset", Hugging Face, revision 4491b06. https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset. CC BY 4.0.

Benchmarks · Epoch AI

Epoch AI, 'Capabilities & benchmarking'. Published online at epoch.ai. Retrieved from 'https://epoch.ai/benchmarks' [online resource]. CC BY 4.0.

Download

Post on X

Text: best place 7

place 65 of 419 on 28 October 2026

1st10th

On the board: 29 January 2026 to 28 October 2026, the newest snapshot
Score: 1445 (1442–1448) on 28 October 2026 · 74,682 votes

WebDev: best place 4

place 85 of 152 on 28 October 2026

1st10th

On the board: 1 February 2026 to 28 October 2026, the newest snapshot
Score: 1436 (1431–1441) on 28 October 2026 · 20,008 votes

Vision: best place 3

place 46 of 167 on 19 October 2026

1st10th

On the board: 29 January 2026 to 19 October 2026, the newest snapshot
Score: 1269 (1263–1275) on 19 October 2026 · 28,249 votes
Matched: gemini-2.5-pro on 29 January 2026, 219 days after it took first place

Epoch AI: Kimi K2.5

released 27 January 2026 · Open weights (unrestricted)

FrontierMath-2025-02-28-Private 27.9%
FrontierMath-Tier-4-2025-07-01-Private 4.2%
GPQA diamond 87.6%
OTIS Mock AIME 2024-2025 92.2%
SWE-Bench verified 73.8%
SimpleQA Verified 34.3%

When each vendor retires it

every model’s dates

Azure AI Foundry 26 Jan 2027 Kimi-K2.5 (version 1)

Sources

Arena scores come from people voting between two anonymous answers. First place is the highest score; the top two are often within each other’s margin of error. LMArena, Arena leaderboard dataset, revision 4491b06, retrieved 30 October 2026; Epoch AI, Capabilities & benchmarking, retrieved 30 October 2026. Both CC BY 4.0. Retirement dates: each vendor’s own deprecation page, quoted and kept, listed with the record.