Kimi-K2-Instruct-0905: benchmark scores, price and ranking

Kimi-K2-Instruct-0905 from Moonshot ranks #258 of 743 LLMs on the LLMs Tiger aggregate benchmark score (49.5), based on 2 benchmark results captured up to 2026-10-13.

Organisation: Moonshot
Type: Open-weight / local model
Open weights: Yes
Released: 2025-09-05 (source: epoch)
Cheapest API price: $0.500 in / $2.00 out per million tokens (via deepinfra, observed 2026-10-13T20:47Z)
Aggregate rank: #258 of 743 (score 49.5)

Across the 2 benchmarks where at least ten models are scored, Kimi-K2-Instruct-0905 is in the top 10% on 0 and in the top half on 1.

Its cheapest listed input price, $0.500 per million tokens, is lower than that of 48% of the 427 models with a listed price.

Kimi-K2-Instruct-0905 benchmark results

Kimi-K2-Instruct-0905 benchmark scores, ranked against all models measured on each benchmark
Benchmark	Score	Rank	Source	Captured
Fiction.liveBench	66.7%	#17 of 47	Epoch AI Benchmarking Hub	2026-10-13
WeirdML	36.7%	#75 of 109	Epoch AI Benchmarking Hub	2026-10-13

About Kimi-K2-Instruct-0905

Kimi-K2-Instruct-0905 is an open-weight model that can run on your own hardware from Moonshot. It was released on 2025-09-05, according to Epoch AI. The cheapest API listing we track charges $0.500 per million input tokens and $2.00 per million output tokens, through deepinfra. We hold 2 benchmark results for it across 1 area, and none is in the top tenth of its board.

Where Kimi-K2-Instruct-0905 stands

On our aggregate score, Kimi-K2-Instruct-0905 is #258 of 743 ranked models, around the middle of the table. Its score is 49.5; the scale is centred on 50, so it sits 0.5 points below the typical model. That beats 65% of the ranked field. Among the 13 ranked Moonshot models it is #6. The next one up is Kimi K2 Thinking (Together) at 49.7. The next one down is Kimi K2.5 (instant) at 49.1.

Kimi-K2-Instruct-0905, benchmark by benchmark

Fiction.liveBench puts Kimi-K2-Instruct-0905 at 66.7%, #17 of 47. That beats 64% of the models tested, but 27.7 points short of grok-4-0709 (94.4%). The results either side are Gemini 2.5 Pro Preview (Mar 2025) (66.7%) and qwen-max-2025-01-25 (66.7%). It covers long-context story comprehension (16k). The board median is 61.1%, so Kimi-K2-Instruct-0905 is 5.6 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

Kimi-K2-Instruct-0905 lands below the middle on WeirdML: 36.7%, #75 of 109, with 68% of the field scoring higher. GPT-6 Astra (max) leads at 93.3%; the closest result above is MiniMax-M2.7 at 37.0%. It covers unusual machine-learning tasks. The board median is 41.2%, so Kimi-K2-Instruct-0905 is 4.5 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

Strengths and weaknesses

General knowledge and chat: uneven, ahead of 48% of the field on average across 2 benchmarks. Its best board here is Fiction.liveBench (#17 of 47) and its weakest is WeirdML (#75 of 109).

Price and value

The median input price across the 427 priced models is $0.500 per million tokens, so Kimi-K2-Instruct-0905 costs about the same as the typical listing. Models that score within 1.5 points or better and cost less: Qwen2.5-Coder-7B-Instruct ($0.010, 49.5), DeepSeek V4.1 Flash (high) ($0.010, 48.9), DeepSeek V4 Flash 0731 (max) ($0.016, 53.1). Similar scores cost more on MiniMax-M1-80k ($0.550) and Kimi K2 Thinking (Together) ($0.550).

Release history

The Moonshot model released before it was Kimi K2 0905 (Novita) on 2025-09-05, which scores 47.3. The next one was Kimi K2.5 on 2026-01-27, at 50.7.

Reading these numbers

The 2 scores on this page come from 1 source: Epoch AI Benchmarking Hub (2). No official leaderboard supplied any of them, so treat them as third-party measurements. With 2 results the aggregate rank can move a lot when one new score arrives.

Models ranked nearby

#255 DeepSeek v4 Flash (max) (49.6)
#256 gpt-4o-2024-11-20 (49.6)
#257 claude-haiku-4-5-20251001_16K (49.5)
#259 mimo-v2.5-pro (49.5)
#260 Reka Flash 3 (49.5)
#261 gemini-exp-1121 (49.5)