vs
3:9
overall capability wins (1 ties across 13 scored). Kimi K2.7 Code leads overall.
- Reasoning1:5
- Coding1:1
- Tool use0:2
- Pricing3:0
- Speed2:0
- Specs1:1
Verdict
Kimi K2.7 Code edges reasoning & knowledge; Kimi K2.7 Code offers larger context (262K vs 256K); Kimi K2.5 allows ~2.0x more max output (64K vs 33K).
- Kimi K2.7 Code edges reasoning & knowledge
- Kimi K2.7 Code offers larger context (262K vs 256K)
- Kimi K2.5 allows ~2.0x more max output (64K vs 33K)
- Kimi K2.5 is ~1.6x cheaper on a blended token basis than Kimi K2.7 Code
Cost for 1M in + 250K out
Illustrative chat workload at primary-provider list prices.
- Kimi K2.5
- $1.23
- Kimi K2.7 Code
- $1.95
Kimi K2.5 is about 1.6x cheaper on this workload.
Price per 1M tokens
Full comparison
Organization
Moonshot
Moonshot
Family
Kimi
Kimi
License
Modified MIT
Modified MIT
Open weights
Yes
Yes
Release
Jan 27, 2026
Jun 12, 2026
Knowledge cutoff
-
-
API / provider
Moonshot
Moonshot
Modalities
text, image → text
text, image, video → text
Specs
Context window
256K
262K
Max output
64K
33K
Parameters
1.0T (32B act.)
1.0T (32B act.)
Pricing
Input $/1M
$0.60
$0.95
Output $/1M
$2.50
$4
Blended $/1M (3∶1)
$1.07
$1.71
Speed
tok/s
70
55
TTFT (s)
0.45
0.5
Reasoning
MMLU-Pro
87.1
—
GPQA Diamond
87.9
89.6
Humanity's Last Exam
30.7
35.0
AIME 2025
96.1
—
MATH-500
—
—
Humanity's Last Exam (with tools)
—
—
AA-Omniscience Accuracy
35.2
39.6
AA-LCR v1.1
78.0
79.3
CritPt
3.1
10.0
MMMU-Pro
75.4
—
IFBench
70.2
63.1
Chartography
—
—
Chartography (With Tools)
—
—
Coding
SWE-bench Verified
76.8
—
SWE-bench Pro
50.7
—
SWE-bench Multilingual
67.3
—
LiveCodeBench
85.0
—
Terminal-Bench 2.1
45.7
67.4
Aider Polyglot
—
—
Terminal-Bench 3
—
—
BigCodeBench
—
—
SciCode
49.0
47.8
CursorBench
—
49.7
SWE-Rebench
49.0
—
NL2Repo-Bench
—
—
DeepSWE
—
—
WebDev Arena
—
1,473
Terminal-Bench 4.0
—
—
CursorBench 4.0
—
—
FrontierCode v1.1 (Main)
—
—
Arena
LMArena Elo
—
—
| Metric | Kimi K2.5 | Kimi K2.7 Code | Delta |
|---|---|---|---|
| Identity | |||
| Organization | Moonshot | Moonshot | — |
| Family | Kimi | Kimi | — |
| License | Modified MIT | Modified MIT | — |
| Open weights | Yes | Yes | — |
| Release | Jan 27, 2026 | Jun 12, 2026 | — |
| Knowledge cutoff | - | - | — |
| API / provider | Moonshot | Moonshot | — |
| Modalities | text, image → text | text, image, video → text | — |
| Specs | |||
| Context window | 256K | 262K | B +6K |
| Max output | 64K | 33K | A +31K |
| Parameters | 1.0T (32B act.) | 1.0T (32B act.) | tie |
| Pricing | |||
| Input $/1M | $0.60 | $0.95 | A +$0.35 |
| Output $/1M | $2.50 | $4 | A +$1.50 |
| Blended $/1M (3∶1) | $1.07 | $1.71 | A +$0.64 |
| Speed | |||
| tok/s | 70 | 55 | A +15 |
| TTFT (s) | 0.45 | 0.5 | A +0.05 |
| Reasoning | |||
| MMLU-Pro | 87.1 | — | — |
| GPQA Diamond | 87.9 | 89.6 | B +1.7 pts |
| Humanity's Last Exam | 30.7 | 35.0 | B +4.3 pts |
| AIME 2025 | 96.1 | — | — |
| MATH-500 | — | — | — |
| Humanity's Last Exam (with tools) | — | — | — |
| AA-Omniscience Accuracy | 35.2 | 39.6 | B +4.4 pts |
| AA-LCR v1.1 | 78.0 | 79.3 | B +1.3 pts |
| CritPt | 3.1 | 10.0 | B +6.9 pts |
| MMMU-Pro | 75.4 | — | — |
| IFBench | 70.2 | 63.1 | A +7.1 pts |
| Chartography | — | — | — |
| Chartography (With Tools) | — | — | — |
| Coding | |||
| SWE-bench Verified | 76.8 | — | — |
| SWE-bench Pro | 50.7 | — | — |
| SWE-bench Multilingual | 67.3 | — | — |
| LiveCodeBench | 85.0 | — | — |
| Terminal-Bench 2.1 | 45.7 | 67.4 | B +21.7 pts |
| Aider Polyglot | — | — | — |
| Terminal-Bench 3 | — | — | — |
| BigCodeBench | — | — | — |
| SciCode | 49.0 | 47.8 | A +1.2 pts |
| CursorBench | — | 49.7 | — |
| SWE-Rebench | 49.0 | — | — |
| NL2Repo-Bench | — | — | — |
| DeepSWE | — | — | — |
| WebDev Arena | — | 1,473 | — |
| Terminal-Bench 4.0 | — | — | — |
| CursorBench 4.0 | — | — | — |
| FrontierCode v1.1 (Main) | — | — | — |
| Arena | |||
| LMArena Elo | — | — | — |
Kimi K2.5
- Context
- 256K / 64K out
- Parameters
- 1.0T (32B act.)
- Price
- $0.60 / $2.50
- Speed
- 70 tok/s · 0.45s TTFT
- Modalities
- text, image → text
- License
- Modified MIT
Prior Kimi open MoE generation - strong long-context agentic coding before K3.
Kimi K2.7 Code
- Context
- 262K / 33K out
- Parameters
- 1.0T (32B act.)
- Price
- $0.95 / $4
- Speed
- 55 tok/s · 0.5s TTFT
- Modalities
- text, image, video → text
- License
- Modified MIT
Open-weight coding specialist (1T MoE / 32B active) with text/image/video input — long-horizon agentic software engineering at $0.95/$4; HighSpeed SKU is $1.90/$8 at ~180 tok/s.
Benchmark charts
Winner bars are emphasized. Per-benchmark deltas sit above each chart.