
Kimi K2.7 Code
Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking
- Context
- 262K
- Released
- 4mo ago
- Parameters
- 1T
- Input / 1M tokens
- $0.67
- Output / 1M tokens
- $3.35
- Cached input / 1M tokens
- $0.18
Sources:models.devOpenRouter
Benchmarks
DeepSWE is the primary coding benchmark where an exact model evaluation exists. Other sourced evaluations appear below; missing scores are not zero.
DeepSWE 1.1
Primary coding-agent benchmark · 113 tasks · Data generated 2026-09-22
- pass@4
- 61.1%
- Average cost / attempt
- $2.82
- Benchmark runs
- 4
pass@1 is the success rate across scored attempts; pass@4 is the share of tasks solved in at least one of four attempts. Costs are reported by DeepSWE for each configuration. These results are not directly comparable to evaluations using different harnesses. Official source JSON.
Other published benchmarks
| Evaluation | Reported score and source |
|---|---|
| FrontierMath-Tier-4-v2-Private | |
| FrontierMath-Tiers-1-3-v2-Private | |
| GPQA diamond | |
| OTIS Mock AIME 2024-2025 | |
| SimpleQA Verified |
Benchmark compilation: Epoch AI, Capabilities & benchmarking (CC BY attribution). Scores are converted to percentages and model names matched to the Vertical Tabs catalog. Individual evaluations are credited to their original sources. External results may carry separate licences and evaluation configurations; do not interpret scores from different harnesses as directly comparable.
Specifications
- Developer
- Moonshot AI
- Model family
- Kimi
- Type
- Language model
- Predecessor
- —
- Successor
- —
- Released
- June 12, 2026
- Model ID
moonshotai/kimi-k2.7-code- Openness
- Open weights
- License
- —
- Parameters
- 1T
- Architecture
- kimi-k2
- Families
- kimi-k2
- Input modalities
- text, image
- Output modalities
- text
- Modality
- text+image->text
- Native context
- 262K
- Tokenizer
- Other
- Max output tokens
- 236K
- Knowledge cutoff
- 2025-01
- Hugging Face
- moonshotai/Kimi-K2.7-Code
- Source freshness
- Metadata synced Oct 8, 2026, 10:17 PM UTC
Open-weight builds
Ollama variants
Every published Ollama tag matched to this model, including parameter editions, instruction/base variants, and quantized builds.
| Variant | Parameters | Type | Quantization | Size | Context | Input |
|---|---|---|---|---|---|---|
| kimi-k2.7-code:cloud | 1T | cloud | — | — | 256K | text + image |
Additional source evidence
- Source model
- moonshotai/kimi-k2.7-code
- Source family
- kimi-k2
- Last updated
- 2026-06-12
- Knowledge cutoff
- 2025-01
- Model context
- 262K
- Model output limit
- 262K
- Open weights
- Yes
- File attachments
- Yes
- Reasoning
- Yes
- Tool calling
- Yes
- Structured output
- Yes
- Temperature control
- No
Downloadable weights
Published benchmarks
| Benchmark | Score | Evaluation details | Evidence |
|---|---|---|---|
| Kimi Code Bench | 62 | Kimi Code CLI · v2 · 2026-06-12 | Source |
| Program Bench | 53.6 | Kimi Code CLI · 2026-06-12 | Source |
| MLS Bench Lite | 35.1 | Kimi Code CLI · 2026-06-12 | Source |
| MCP Atlas | 76 · success rate | Kimi Code CLI · 2026-06-12 | Source |
| MCP Mark Verified | 81.1 · success rate | Kimi Code CLI · 2026-06-12 | Source |
| Kimi Claw 24/7 Bench | 46.9 | Kimi Code CLI · 2026-06-12 | Source |
Complete source record (JSON)
{
"id": "moonshotai/kimi-k2.7-code",
"name": "Kimi K2.7 Code",
"description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking",
"family": "kimi-k2",
"last_updated": "2026-06-12",
"knowledge": "2025-01",
"attachment": true,
"reasoning": true,
"tool_call": true,
"structured_output": true,
"temperature": false,
"modalities": {
"input": [
"text",
"image",
"video"
],
"output": [
"text"
]
},
"limit": {
"context": 262144,
"output": 262144
},
"benchmarks": [
{
"name": "Kimi Code Bench",
"score": 62,
"source": "https://huggingface.co/moonshotai/Kimi-K2.7-Code",
"harness": "Kimi Code CLI",
"version": "v2",
"date": "2026-06-12"
},
{
"name": "Program Bench",
"score": 53.6,
"source": "https://huggingface.co/moonshotai/Kimi-K2.7-Code",
"harness": "Kimi Code CLI",
"date": "2026-06-12"
},
{
"name": "MLS Bench Lite",
"score": 35.1,
"source": "https://huggingface.co/moonshotai/Kimi-K2.7-Code",
"harness": "Kimi Code CLI",
"date": "2026-06-12"
},
{
"name": "MCP Atlas",
"score": 76,
"source": "https://huggingface.co/moonshotai/Kimi-K2.7-Code",
"metric": "success rate",
"harness": "Kimi Code CLI",
"date": "2026-06-12"
},
{
"name": "MCP Mark Verified",
"score": 81.1,
"source": "https://huggingface.co/moonshotai/Kimi-K2.7-Code",
"metric": "success rate",
"harness": "Kimi Code CLI",
"date": "2026-06-12"
},
{
"name": "Kimi Claw 24/7 Bench",
"score": 46.9,
"source": "https://huggingface.co/moonshotai/Kimi-K2.7-Code",
"harness": "Kimi Code CLI",
"date": "2026-06-12"
}
],
"release_date": "2026-06-12",
"open_weights": true,
"weights": [
{
"label": "Hugging Face",
"url": "https://huggingface.co/moonshotai/Kimi-K2.7-Code"
}
]
}Infrastructure
Provider availability
| Provider | Endpoint | Context | Input /M tokens | Output /M tokens | Cached input /M tokens | Cache write /M tokens | Uptime 1d | Data handling |
|---|---|---|---|---|---|---|---|---|
| inceptron/int4 | 262K | $0.67 | $3.35 | $0.18 | — | 99.86% | ZDR: YesTraining: Unknown Prompt logging: Unknown Retention: Unknown Checked 2026-10-15 | |
| coreweave/int4 | 262K | $0.71 | $3.50 | $0.15 | — | 100.00% | ZDR: YesTraining: Unknown Prompt logging: Unknown Retention: Unknown Checked 2026-10-15 | |
| streamlake | 256K | $0.71 | $3.00 | $0.14 | — | 99.75% | ZDR: NoTraining: Unknown Prompt logging: Unknown Retention: Unknown Checked 2026-10-15 | |
| venice/int4 | 256K | $0.75 | $3.50 | $0.16 | — | 100.00% | ZDR: YesTraining: Unknown Prompt logging: Unknown Retention: Unknown Checked 2026-10-15 | |
| siliconflow/fp8 | 262K | $0.86 | $3.80 | $0.18 | — | 100.00% | ZDR: YesTraining: Unknown Prompt logging: Unknown Retention: Unknown Checked 2026-10-15 | |
| novita/int4 | 262K | $0.91 | $3.84 | $0.18 | — | 100.00% | ZDR: YesTraining: Unknown Prompt logging: Unknown Retention: Unknown Checked 2026-10-15 | |
| alibaba/fp8 | 262K | $0.95 | $4.00 | $0.19 | — | 99.17% | ZDR: NoTraining: Unknown Prompt logging: Unknown Retention: Unknown Checked 2026-10-15 | |
| cloudflare | 262K | $0.95 | $4.00 | $0.19 | — | 100.00% | ZDR: NoTraining: Unknown Prompt logging: Unknown Retention: Unknown Checked 2026-10-15 | |
| gmicloud/fp8 | 262K | $0.95 | $4.00 | $0.19 | — | 86.67% | ZDR: NoTraining: Unknown Prompt logging: Unknown Retention: Unknown Checked 2026-10-15 | |
| moonshotai/int4 | 262K | $0.95 | $4.00 | $0.19 | — | 100.00% | ZDR: YesTraining: Unknown Prompt logging: Unknown Retention: Unknown Checked 2026-10-15 | |
| nebius/fp4 | 262K | $0.95 | $4.00 | — | — | 100.00% | ZDR: YesTraining: Unknown Prompt logging: Unknown Retention: Unknown Checked 2026-10-15 | |
| phala | 262K | $0.95 | $4.00 | $0.19 | — | 98.69% | ZDR: YesTraining: Unknown Prompt logging: Unknown Retention: Unknown Checked 2026-10-15 | |
| moonshotai/highspeed | 262K | $1.90 | $8.00 | $0.38 | — | 100.00% | ZDR: YesTraining: Unknown Prompt logging: Unknown Retention: Unknown Checked 2026-10-15 |
Provider data:OpenRouter
Deployment graph
Can I run this?
Vertical Tabs only shows compatibility when it has structured evidence tying this exact model to hardware, runtime, quantization, memory use, and observed performance.