in the loop

Kimi K2.7 Code
#55 of 274 scored.

MoonshotJun 12, 2026Open weights

Capabilities Index

Range 147.9–151.5

150.0

Context

Takes text, image

262K

Input

Per 1M tokens

$0.67

Output

Per 1M tokens

$3.35

Benchmarks

How it scores.

  • GPQA Diamond83.8%

    49th of 186

  • FrontierMath54.0%

    43rd of 81

  • SimpleQA Verified36.5%

    48th of 77

  • Mock AIME95.5%

    32nd of 176

Everything else Epoch has for it

  • WeirdML54.1%
  • SimpleBench49.5%
  • Surface Evolver Bench48.8%
  • APEX-Agents37.6%
  • DeepSWE30.5%
  • FrontierCode30.1%
  • Chess Puzzles16.9%
  • FrontierMath-Tier-4-v2-Private12.2%

On the index

Around it.

  1. #52Inkling-SmallThinking Machines · $0.45 in150.2
  2. #53Claude Opus 4.5Anthropic · $5.00 in150.1
  3. #54GPT-5OpenAI · $1.25 in150.0
  4. #55Kimi K2.7 CodeMoonshot · $0.67 in150.0
  5. #56GLM-5.1Z.ai · $0.97 in149.8
  6. #57GPT-5.1OpenAI · $1.25 in149.6
  7. #58Qwen 3.8 27BAlibaba · $0.42 in149.4

Sources