Kimi K2.6
Multimodal Kimi workhorse for agent loops, coding tasks, and visual context
- Strongest at
- Writing Better than 49% of the past year's models · 4 of 4 benchmarks
- Price per 1M tokens
- $0.95 / $4 input / output · Moonshot AI
- Context window
- 262.1K tokens in · 262.1K out
- Since the last one
- 83 days after Kimi K2.5
Capability profile
Scored against 91 models released in the past year. 50 is the average model; further out is better.
Where these scores come from
Each capability combines the published benchmarks listed here, rescaled so 50 is the reference average and 15 points is one standard deviation. A hollow dot rests on a single benchmark.
- Coding SWE-Bench verified, LiveBench: Coding + Agentic Coding
- Writing EQ-Bench Creative Writing v3, EQ-Bench Longform Writing, Lech Mazur Writing comparison, LiveBench: Instruction Following
- Reading & documents GDP.pdf (Surge AI), Vision Evals: OCR & extraction (low effort), LiveBench: Language + Data Analysis
- Reasoning Chess Puzzles, Mystery Game Puzzles, LiveBench: Reasoning
- Math OTIS Mock AIME 2024-2025, FrontierMath-Tiers-1-3-v2-Private, FrontierMath-Tier-4-v2-Private, ProofBench, LiveBench: Mathematics
- Science GPQA diamond, WeirdML v2
- Visual reasoning Vision Evals (low effort), Blueprint-Bench 2, GDP.pdf (Surge AI)
- Computer use AutomationBench 1.0.6
Price against capability
The shaded corner holds anything better and cheaper. Prices blend input and output 3 to 1.
11 models from the past year score higher on Writing for less: Gemini 3.8 Flash, Gemini 3.6 Flash, Kimi K2.5 and 8 more.
11 models from the past year score higher on Coding for less: DeepSeek V4.1 Flash, DeepSeek V3.2, Claude Haiku 5.5 and 8 more.
21 models from the past year score higher on Reading & documents for less: Gemini 3 Flash Preview, Qwen3.5 9B, DeepSeek V4 Pro 0813 and 18 more.
20 models from the past year score higher on Reasoning for less: Gemini 3.8 Flash, DeepSeek V4.1 Flash 1017, Gemini 3.7 Flash and 17 more.
19 models from the past year score higher on Math for less: Claude Haiku 5.5, DeepSeek V3.2, GPT-6 Luna and 16 more.
17 models from the past year score higher on Science for less: Gemini 3.8 Flash, DeepSeek V4.1 Flash 1017, GPT-6 Luna and 14 more.
15 models from the past year score higher on Visual reasoning for less: Gemini 3.7 Flash, Gemini 3.8 Flash, Claude Haiku 5.5 and 12 more.
13 models from the past year score higher on Computer use for less: Claude Haiku 5.5, Qwen3.8 27B, DeepSeek V4 Pro 0813 and 10 more.
OpenRouter $0.65 / $3.41 · Moonshot AI $0.95 / $4 · Cached input $0.16
Since Kimi K2.5
5 releases in the Kimi line.
A newer release is out: Kimi K2.8 Preview.
Epoch benchmark boards
Best result on each board it has run. The tick marks the leader.
7 more boards
Moonshot AI's past year
7 release days in the past year, about one every 52 days. The next is expected around Dec 10, 2026. Forecasts ↗
Specifications
Specifications and prices from Models.dev; availability and limits may differ by provider. Benchmarks from Epoch AI (CC BY 4.0), LiveBench and the other sources named above.