Kimi K2: benchmark scores, price and ranking

Kimi K2 from Moonshot AI ranks #98 of 743 LLMs on the LLMs Tiger aggregate benchmark score (51.8), based on 2 benchmark results captured up to 2026-10-13.

Organisation: Moonshot AI
Type: Open-weight / local model
Open weights: Yes
Released: 2025-07-11 (source: curated)
Cheapest API price: $0.550 in / $2.20 out per million tokens (via vercel_ai_gateway, observed 2026-10-13T20:47Z)
Aggregate rank: #98 of 743 (score 51.8)

Across the 2 benchmarks where at least ten models are scored, Kimi K2 is in the top 10% on 0 and in the top half on 2.

Its cheapest listed input price, $0.550 per million tokens, is lower than that of 47% of the 427 models with a listed price.

Kimi K2 benchmark results

Kimi K2 benchmark scores, ranked against all models measured on each benchmark
Benchmark	Score	Rank	Source	Captured
Aider Polyglot	59.1%	#13 of 49	Aider polyglot leaderboard	2026-10-13
SWE-bench Verified	65.4%	#26 of 73	SWE-bench Verified leaderboard	2026-10-13

About Kimi K2

Kimi K2 is an open-weight model that can run on your own hardware from Moonshot AI. It was released on 2025-07-11, according to the developer's announcement. The cheapest API listing we track charges $0.550 per million input tokens and $2.20 per million output tokens, through vercel_ai_gateway. We hold 2 benchmark results for it across 1 area, and none is in the top tenth of its board.

Where Kimi K2 stands

On our aggregate score, Kimi K2 is #98 of 743 ranked models, in the top quarter. Its score is 51.8; the scale is centred on 50, so it sits 1.8 points above the typical model. That beats 87% of the ranked field. Among the 3 ranked Moonshot AI models it is #1. The next one down is kimi-k2p5 at 49.2.

Kimi K2, benchmark by benchmark

Aider Polyglot puts Kimi K2 at 59.1%, #13 of 49. That beats 73% of the models tested, but 24.0 points short of gemini-2.5-pro-preview-06-05 (83.1%). The results either side are Qwen3 235B A22B diff, no think, Alibaba API (59.6%) and DeepSeek-R1 (56.9%). It covers edits across 6 languages (225 exercises). The board median is 44.9%, so Kimi K2 is 14.2 points above it. Source: Aider polyglot leaderboard, captured 2026-10-13.

On SWE-bench Verified, Kimi K2 reaches 65.4%. It ranks #26 of 73, ahead of 64% of the field, and 18.1 points behind Claude Opus 4.7 (max) at 83.5%. The results either side are GPT-5.1 Codex (66.0%) and o1-preview (64.6%). It covers real GitHub issues resolved. The board median is 56.4%, so Kimi K2 is 9.0 points above it. Source: SWE-bench Verified leaderboard, captured 2026-10-13.

Strengths and weaknesses

Software engineering: solid, ahead of 69% of the field on average across 2 benchmarks. Its best board here is Aider Polyglot (#13 of 49) and its weakest is SWE-bench Verified (#26 of 73).

Price and value

The median input price across the 427 priced models is $0.500 per million tokens, so Kimi K2 costs more than the typical listing. Models that score within 1.5 points or better and cost less: DeepSeek V4 Flash 0731 (max) ($0.016, 53.1), Qwen3.7 Flash ($0.030, 50.5), GLM-5.2 (high) ($0.060, 51.6). Similar scores cost more on DeepSeek-V3.2-Speciale ($0.580) and GLM-5 ($0.600).

Reading these numbers

The 2 scores on this page come from 2 sources: Aider polyglot leaderboard (1), SWE-bench Verified leaderboard (1). 2 of them come from an official leaderboard and 0 from datasets or third-party runs. With 2 results the aggregate rank can move a lot when one new score arrives.

Models ranked nearby

#95 DeepSeek R1 + claude-3-5-sonnet-20241022 (51.8)
#96 Qwen3.8-Flash-Next (51.8)
#97 GPT-4 (0806) (51.8)
#99 Ornith-1.5-35B-A3B (51.8)
#100 MiMo-V2.6 Flash (51.7)
#101 DeepSeek-V2.5-1210 (51.7)