Kimi K2.7 Code: benchmarks, price and rank
Kimi K2.7 Code from Moonshot is 68th of 123 rated AI models on Spezia's leaderboard. It costs $0.68 in, $3.40 out per million tokens.
- 68th of 123 rated models
- Rating 1,170
- Rated on 9 benchmarks
- Open weights
- Released Jun 12, 2026
- 262,000 tokens of context
Where it is strongest
Not in the top tenth on any benchmark in use.
Where it falls behind
- 57th of 66 on FrontierMath Tier 4
- 47th of 56 on APEX-Agents
- 52nd of 76 on FrontierMath
Price: $0.68 in, $3.40 out per million tokens.
Its headline results
The benchmarks that count most in the rating, with its place among the models measured the same way, and where each number came from.
- APEX-Agents37.6%
Long tasks written by investment banking analysts, management consultants and corporate lawyers from their own work, done across files and apps in a realistic work setup.
47th of 56. Independent result from Epoch AI.
- MCP Atlas76%
Multi-step tasks where the model has to pick and combine many real tools, connected through MCP, the standard way agents plug into apps and data.
11th of 33. The maker's own number, from Moonshot AI, via LLM Stats.
- Vending-Bench 2$5,083
The model runs a simulated vending machine business for a year on its own: finding suppliers, ordering stock, setting prices and paying the daily fees.
35th of 60. Independent result from Epoch AI.
- DeepSWE 1.131%
Software engineering tasks run in the same simple agent setup for every model.
Independent result from LLM Stats.
- SimpleBench57.9%
Multiple-choice puzzles about everyday situations, space, time and people, written to trip up models that pattern-match instead of think. Ordinary adults do well on them.
46th of 68. Independent result from Epoch AI.
- SimpleQA Verified36.5%
Short factual questions with one checked answer, asked without web search.
52nd of 78. Independent result from Epoch AI.
- FrontierMath54.0%
Hundreds of new, unpublished math problems written by mathematicians, from hard undergraduate level up to research, with answers a computer can check.
52nd of 76. Independent result from Epoch AI.
- LMArena WebDev1,474
People ask two anonymous models to build the same web page or app and vote for the better one.
57th of 98. Independent result from LMArena.
- GPQA Diamond87.9%
Graduate-level questions in biology, physics and chemistry, written by PhDs so that searching the web does not give the answer away.
60th of 103. Independent result from Epoch AI.
- FrontierMath Tier 412.2%
Research-level problems that take professional mathematicians days.
57th of 66. Independent result from Epoch AI.
1 more results in its full profile.
Every result with its source, what it costs per task, what jumped with its release, side by side with any other model, and answers to your questions about when to use it.
Models ranked near it
How the rating works, and every benchmark it uses: the model leaderboard.