vs
4:1
overall capability wins (0 ties across 5 scored). Kimi K2.6 leads overall.
- Reasoning4:0
- Specs0:1
Verdict
Kimi K2.6 edges reasoning & knowledge; Kimi Linear 48B A3B Instruct offers ~4.1x larger context (1.0M vs 256K); Kimi Linear 48B A3B Instruct ships open weights (MIT).
- Kimi K2.6 edges reasoning & knowledge
- Kimi Linear 48B A3B Instruct offers ~4.1x larger context (1.0M vs 256K)
- Kimi Linear 48B A3B Instruct ships open weights (MIT)
Cost for 1M in + 250K out
Illustrative chat workload at primary-provider list prices.
- Kimi K2.6
- $1
- Kimi Linear 48B A3B Instruct
- -
Price per 1M tokens
Full comparison
Organization
Moonshot
Moonshot
Family
Kimi
Kimi
License
Proprietary
MIT
Open weights
No
Yes
Release
Apr 8, 2026
Oct 30, 2025
Knowledge cutoff
-
-
API / provider
Moonshot
No primary API price listed
Modalities
text, image → text
text → text
Specs
Context window
256K
1.0M
Max output
16K
-
Parameters
—
48B (3B act.)
Pricing
Input $/1M
$0.50
—
Output $/1M
$2
—
Blended $/1M (3∶1)
$0.88
-
Speed
tok/s
70
-
TTFT (s)
0.4
-
Reasoning
MMLU-Pro
—
—
GPQA Diamond
91.1
41.2
Humanity's Last Exam
37.5
2.5
AIME 2025
—
—
MATH-500
—
—
Humanity's Last Exam (with tools)
—
—
AA-Omniscience Accuracy
32.6
—
AA-LCR v1.1
81.0
28.0
CritPt
8.0
—
MMMU-Pro
79.4
—
IFBench
76.0
28.1
Chartography
—
—
Chartography (With Tools)
—
—
Coding
SWE-bench Verified
—
—
SWE-bench Pro
—
—
SWE-bench Multilingual
—
—
LiveCodeBench
—
—
Terminal-Bench 2.1
65.9
—
Aider Polyglot
—
—
Terminal-Bench 3
—
—
BigCodeBench
—
—
SciCode
51.5
—
CursorBench
—
—
SWE-Rebench
46.5
—
NL2Repo-Bench
—
—
DeepSWE
—
—
WebDev Arena
1,509
—
Terminal-Bench 4.0
—
—
CursorBench 4.0
—
—
FrontierCode v1.1 (Main)
—
—
Arena
LMArena Elo
1,460
—
| Metric | Kimi K2.6 | Kimi Linear 48B A3B Instruct | Delta |
|---|---|---|---|
| Identity | |||
| Organization | Moonshot | Moonshot | — |
| Family | Kimi | Kimi | — |
| License | Proprietary | MIT | — |
| Open weights | No | Yes | — |
| Release | Apr 8, 2026 | Oct 30, 2025 | — |
| Knowledge cutoff | - | - | — |
| API / provider | Moonshot | No primary API price listed | — |
| Modalities | text, image → text | text → text | — |
| Specs | |||
| Context window | 256K | 1.0M | B +793K |
| Max output | 16K | - | — |
| Parameters | — | 48B (3B act.) | — |
| Pricing | |||
| Input $/1M | $0.50 | — | — |
| Output $/1M | $2 | — | — |
| Blended $/1M (3∶1) | $0.88 | - | — |
| Speed | |||
| tok/s | 70 | - | — |
| TTFT (s) | 0.4 | - | — |
| Reasoning | |||
| MMLU-Pro | — | — | — |
| GPQA Diamond | 91.1 | 41.2 | A +49.9 pts |
| Humanity's Last Exam | 37.5 | 2.5 | A +35.0 pts |
| AIME 2025 | — | — | — |
| MATH-500 | — | — | — |
| Humanity's Last Exam (with tools) | — | — | — |
| AA-Omniscience Accuracy | 32.6 | — | — |
| AA-LCR v1.1 | 81.0 | 28.0 | A +53.0 pts |
| CritPt | 8.0 | — | — |
| MMMU-Pro | 79.4 | — | — |
| IFBench | 76.0 | 28.1 | A +47.9 pts |
| Chartography | — | — | — |
| Chartography (With Tools) | — | — | — |
| Coding | |||
| SWE-bench Verified | — | — | — |
| SWE-bench Pro | — | — | — |
| SWE-bench Multilingual | — | — | — |
| LiveCodeBench | — | — | — |
| Terminal-Bench 2.1 | 65.9 | — | — |
| Aider Polyglot | — | — | — |
| Terminal-Bench 3 | — | — | — |
| BigCodeBench | — | — | — |
| SciCode | 51.5 | — | — |
| CursorBench | — | — | — |
| SWE-Rebench | 46.5 | — | — |
| NL2Repo-Bench | — | — | — |
| DeepSWE | — | — | — |
| WebDev Arena | 1,509 | — | — |
| Terminal-Bench 4.0 | — | — | — |
| CursorBench 4.0 | — | — | — |
| FrontierCode v1.1 (Main) | — | — | — |
| Arena | |||
| LMArena Elo | 1,460 | — | — |
Kimi K2.6
- Context
- 256K / 16K out
- Parameters
- —
- Price
- $0.50 / $2
- Speed
- 70 tok/s · 0.4s TTFT
- Modalities
- text, image → text
- License
- Proprietary
Post-K2.5 Moonshot API SKU with longer context and stronger multimodal agent tooling.
Kimi Linear 48B A3B Instruct
- Context
- 1.0M
- Parameters
- 48B (3B act.)
- Price
- —
- Speed
- —
- Modalities
- text → text
- License
- MIT
Moonshot AI's hybrid linear-attention MoE (48B total / 3B active) pairing Kimi Delta Attention with sparse global attention for cheaper long-context (up to 1M token) inference; no hosted API yet.
Benchmark charts
Winner bars are emphasized. Per-benchmark deltas sit above each chart.