File size: 24,750 Bytes
dc4be28
c46f181
 
dc4be28
 
8a112f5
 
 
 
 
b51274b
8a112f5
b51274b
8a112f5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
7f98307
 
 
 
 
8a112f5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b51274b
 
 
 
 
 
 
8a112f5
 
 
 
 
 
 
 
 
 
 
 
 
b51274b
8a112f5
 
 
 
 
 
 
 
 
 
 
 
 
 
b51274b
8a112f5
 
 
b51274b
8a112f5
 
 
 
 
 
 
b51274b
8a112f5
 
7f98307
8a112f5
 
 
 
 
 
 
 
b51274b
8a112f5
 
 
 
 
 
b51274b
8a112f5
 
 
 
b51274b
8a112f5
 
 
 
 
 
 
 
 
 
 
 
 
b51274b
8a112f5
 
 
 
 
 
 
 
 
b51274b
8a112f5
b51274b
8a112f5
b51274b
8a112f5
 
 
 
b51274b
8a112f5
 
b51274b
8a112f5
b51274b
7f98307
8a112f5
 
b51274b
8a112f5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b51274b
8a112f5
 
 
 
 
 
 
 
 
b51274b
8a112f5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b51274b
8a112f5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b51274b
 
8a112f5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b51274b
8a112f5
 
 
 
 
 
7f98307
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
8a112f5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b51274b
8a112f5
 
b51274b
8a112f5
b51274b
8a112f5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b51274b
8a112f5
 
 
 
 
 
 
b51274b
8a112f5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b51274b
 
8a112f5
 
 
 
 
 
 
 
 
 
2f7e011
 
 
 
 
 
8a112f5
 
7f98307
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
---
license: other
license_name: modified-mit
library_name: transformers
---
<div align="center">
  <picture>
      <img src="figures/kimi-logo.png" width="30%" alt="Kimi K2: Open Agentic Intellignece">
  </picture>
</div>

<hr>

<div align="center" style="line-height:1">
  <a href="https://www.kimi.com" target="_blank"><img alt="Chat" src="https://img.shields.io/badge/🤖%20Chat-Kimi%20K2-ff6b6b?color=1783ff&logoColor=white"/></a>
  <a href="https://www.moonshot.ai" target="_blank"><img alt="Homepage" src="https://img.shields.io/badge/Homepage-Moonshot%20AI-white?logo=Kimi&logoColor=white"/></a>
</div>

<div align="center" style="line-height: 1;">
  <a href="https://huggingface.co/moonshotai" target="_blank"><img alt="Hugging Face" src="https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-Moonshot%20AI-ffc107?color=ffc107&logoColor=white"/></a>
  <a href="https://twitter.com/kimi_moonshot" target="_blank"><img alt="Twitter Follow" src="https://img.shields.io/badge/Twitter-Kimi.ai-white?logo=x&logoColor=white"/></a>
    <a href="https://discord.gg/TYU2fdJykW" target="_blank"><img alt="Discord" src="https://img.shields.io/badge/Discord-Kimi.ai-white?logo=discord&logoColor=white"/></a>
</div>

<div align="center" style="line-height: 1;">
  <a href="https://github.com/moonshotai/Kimi-K2/blob/main/LICENSE"><img alt="License" src="https://img.shields.io/badge/License-Modified_MIT-f5de53?&color=f5de53"/></a>
</div>

<p align="center">
<b>📰&nbsp;&nbsp;<a href="https://moonshotai.github.io/Kimi-K2/">Tech Blog</a></b> &nbsp;&nbsp;&nbsp; | &nbsp;&nbsp;&nbsp; <b>📄&nbsp;&nbsp;Paper Link (comming soon)</b>
</p>

## 1. Model Introduction

Kimi K2 is a state-of-the-art mixture-of-experts (MoE) language model with 32 billion activated parameters and 1 trillion total parameters. Trained with the Muon optimizer, Kimi K2 achieves exceptional performance across frontier knowledge, reasoning, and coding tasks while being meticulously optimized for agentic capabilities.

### Key Features
- Large-Scale Training: Pre-trained a 1T parameter MoE model on 15.5T tokens with zero training instability.
- MuonClip Optimizer: We apply the Muon optimizer to an unprecedented scale, and develop novel optimization techniques to resolve instabilities while scaling up.
- Agentic Intelligence: Specifically designed for tool use, reasoning, and autonomous problem-solving.

### Model Variants
- **Kimi-K2-Base**: The foundation model, a strong start for researchers and builders who want full control for fine-tuning and custom solutions.
- **Kimi-K2-Instruct**: The post-trained model best for drop-in, general-purpose chat and agentic experiences. It is a reflex-grade model without long thinking.

<div align="center">
  <picture>
      <img src="figures/banner.png" width="80%" alt="Evaluation Results">
  </picture>
</div>

## 2. Model Summary

<div align="center">


| | |
|:---:|:---:|
| **Architecture** | Mixture-of-Experts (MoE) |
| **Total Parameters** | 1T |
| **Activated Parameters** | 32B |
| **Number of Layers** (Dense layer included) | 61 |
| **Number of Dense Layers** | 1 |
| **Attention Hidden Dimension** | 7168 |
| **MoE Hidden Dimension** (per Expert) | 2048 |
| **Number of Attention Heads** | 64 |
| **Number of Experts** | 384 |
| **Selected Experts per Token** | 8 |
| **Number of Shared Experts** | 1 |
| **Vocabulary Size** | 160K |
| **Context Length** | 128K |
| **Attention Mechanism** | MLA |
| **Activation Function** | SwiGLU |
</div>

## 3. Evaluation Results

#### Instruction model evaluation results

<div align="center">
<table>
<thead>
<tr>
<th align="center">Benchmark</th>
<th align="center">Metric</th>
<th align="center"><sup>Kimi K2 Instruct</sup></th>
<th align="center"><sup>DeepSeek-V3-0324</sup></th>
<th align="center"><sup>Qwen3-235B-A22B <br><sup>(non-thinking)</sup></sup></th>
<th align="center"><sup>Claude Sonnet 4 <br><sup>(w/o extended thinking)</sup></sup></th>
<th align="center"><sup>Claude Opus 4 <br><sup>(w/o extended thinking)</sup></sup></th>
<th align="center"><sup>GPT-4.1</sup></th>
<th align="center"><sup>Gemini 2.5 Flash <br> Preview (05-20)</sup></th>
</tr>
</thead>
<tbody>
<tr>
<td align="center" colspan=9><strong>Coding Tasks</strong></td>
</tr>
<tr>
<td align="center">LiveCodeBench v6<br><sup>(Aug 24 - May 25)</sup></td>
<td align="center">Pass@1</td>
<td align="center"><strong>53.7</strong></td>
<td align="center">46.9</td>
<td align="center">37.0</td>
<td align="center">48.5</td>
<td align="center">47.4</td>
<td align="center">44.7</td>
<td align="center">44.7</td>
</tr>
<tr>
<td align="center">OJBench</td>
<td align="center">Pass@1</td>
<td align="center"><strong>27.1</strong></td>
<td align="center">24.0</td>
<td align="center">11.3</td>
<td align="center">15.3</td>
<td align="center">19.6</td>
<td align="center">19.5</td>
<td align="center">19.5</td>
</tr>

<tr>
<td align="center">MultiPL-E</td>
<td align="center">Pass@1</td>
<td align="center"><ins><strong>85.7</strong></ins></td>
<td align="center">83.1</td>
<td align="center">78.2</td>
<td align="center">88.6</td>
<td align="center"><strong>89.6</strong></td>
<td align="center">86.7</td>
<td align="center">85.6</td>
</tr>

<tr>
<td align="center">SWE-bench Verified <br/><sup>(Agentless Coding)</sup></td>
<td align="center">Single Patch w/o Test (Acc)</td>
<td align="center"><ins><strong>51.8</strong></ins></td>
<td align="center">36.6</td>
<td align="center">39.4</td>
<td align="center">50.2</td>
<td align="center"><strong>53.0</strong></td>
<td align="center">40.8</td>
<td align="center">32.6</td>
</tr>

<tr>
<td align="center" rowspan="2">SWE-bench Verified <br/> <sup>(Agentic Coding)</sup></td>
<td align="center">Single Attempt (Acc)</td>
<td align="center"><ins><strong>65.8</strong></ins></td>
<td align="center">38.8</td>
<td align="center">34.4</td>
<td align="center"><strong>72.7</strong><sup>*</sup></td>
<td align="center">72.5<sup>*</sup></td>
<td align="center">54.6</td>
<td align="center"></td>
</tr>

<tr>
<!--<td align="center">(Agentic Coding)</td>-->
<td align="center">Multiple Attempts (Acc)</td>
<td align="center"><ins><strong>71.6</strong></ins></td>
<td align="center"></td>
<td align="center"></td>
<td align="center"><strong>80.2</strong></td>
<td align="center">79.4<sup>*</sup></td>
<td align="center"></td>
<td align="center"></td>
</tr>

<tr>
<td align="center">SWE-bench Multilingual<br /> <sup>(Agentic Coding)</sup></td>
<td align="center">Single Attempt (Acc)</td>
<td align="center"><ins><strong>47.3</strong> </ins></td>
<td align="center">25.8</td>
<td align="center">20.9</td>
<td align="center"><strong>51.0</strong></td>
<td align="center"></td>
<td align="center">31.5</td>
<td align="center"></td>
</tr>

<tr>
<td align="center" rowspan="2">TerminalBench</td>
<td align="center">Inhouse Framework (Acc)</td>
<td align="center"><ins><strong>30.0</strong></ins></td>
<td align="center"></td>
<td align="center"></td>
<td align="center">35.5</td>
<td align="center"><strong>43.2</strong></td>
<td align="center">8.3</td>
<td align="center"></td>
</tr>

<tr>
<!--<td align="center">TerminalBench</td>-->
<td align="center">Terminus (Acc)</td>
<td align="center"><ins><strong>25.0</strong> </ins></td>
<td align="center">16.3</td>
<td align="center">6.6</td>
<td align="center"></td>
<td align="center"></td>
<td align="center"><strong>30.3</strong></td>
<td align="center">16.8</td>
</tr>
<tr>
<td align="center">Aider-Polyglot</td>
<td align="center">Acc</td>
<td align="center">60.0</td>
<td align="center">55.1</td>
<td align="center"><ins><strong>61.8</strong></ins></td>
<td align="center">56.4</td>
<td align="center"><strong>70.7</strong></td>
<td align="center">52.4</td>
<td align="center">44.0</td>
</tr>
<tr>
<td align="center" colspan=9><strong>Tool Use Tasks</strong></td>
</tr>
<tr>
<td align="center">Tau2 retail</td>
<td align="center">Avg@4</td>
<td align="center"><ins><strong>70.6</strong></ins></td>
<td align="center">69.1</td>
<td align="center">57.0</td>
<td align="center">75.0</td>
<td align="center"><strong>81.8</strong></td>
<td align="center">74.8</td>
<td align="center">64.3</td>
</tr>
<tr>
<td align="center">Tau2 airline</td>
<td align="center">Avg@4</td>
<td align="center"><ins><strong>56.5</strong></ins></td>
<td align="center">39.0</td>
<td align="center">26.5</td>
<td align="center">55.5</td>
<td align="center"><strong>60.0</strong></td>
<td align="center">54.5</td>
<td align="center">42.5</td>
</tr>
<tr>
<td align="center">Tau2 telecom</td>
<td align="center">Avg@4</td>
<td align="center"><strong>65.8</strong></td>
<td align="center">32.5</td>
<td align="center">22.1</td>
<td align="center">45.2</td>
<td align="center">57.0</td>
<td align="center">38.6</td>
<td align="center">16.9</td>
</tr>
<tr>
<td align="center">AceBench</td>
<td align="center">Acc</td>
<td align="center"><ins><strong>76.5</strong></ins></td>
<td align="center">72.7</td>
<td align="center">70.5</td>
<td align="center">76.2</td>
<td align="center">75.6</td>
<td align="center"><strong>80.1</strong></td>
<td align="center">74.5</td>
</tr>
<tr>
<td align="center" colspan=9><strong>Math &amp; STEM Tasks</strong></td>
</tr>
<tr>
<td align="center">AIME 2024</td>
<td align="center">Avg@64</td>
<td align="center"><strong>69.6</strong></td>
<td align="center">59.4<sup>*</sup></td>
<td align="center">40.1<sup>*</sup></td>
<td align="center">43.4</td>
<td align="center">48.2</td>
<td align="center">46.5</td>
<td align="center">61.3</td>
</tr>
<tr>
<td align="center">AIME 2025</td>
<td align="center">Avg@64</td>
<td align="center"><strong>49.5</strong></td>
<td align="center">46.7</td>
<td align="center">24.7<sup>*</sup></td>
<td align="center">33.1<sup>*</sup></td>
<td align="center">33.9<sup>*</sup></td>
<td align="center">37.0</td>
<td align="center">46.6</td>
</tr>
<tr>
<td align="center">MATH-500</td>
<td align="center">Acc</td>
<td align="center"><strong>97.4</strong></td>
<td align="center">94.0<sup>*</sup></td>
<td align="center">91.2<sup>*</sup></td>
<td align="center">94.0</td>
<td align="center">94.4</td>
<td align="center">92.4</td>
<td align="center">95.4</td>
</tr>
<tr>
<td align="center">HMMT 2025</td>
<td align="center">Avg@32</td>
<td align="center"><strong>38.8</strong></td>
<td align="center">27.5</td>
<td align="center">11.9</td>
<td align="center">15.9</td>
<td align="center">15.9</td>
<td align="center">19.4</td>
<td align="center">34.7</td>
</tr>
<tr>
<td align="center">CNMO 2024</td>
<td align="center">Avg@16</td>
<td align="center">74.3</td>
<td align="center"><ins><strong>74.7</strong></ins></td>
<td align="center">48.6</td>
<td align="center">60.4</td>
<td align="center">57.6</td>
<td align="center">56.6</td>
<td align="center"><strong>75.0</strong></td>
</tr>
<tr>
<td align="center">PolyMath-en</td>
<td align="center">Avg@4</td>
<td align="center"><strong>65.1</strong></td>
<td align="center">59.5</td>
<td align="center">51.9</td>
<td align="center">52.8</td>
<td align="center">49.8</td>
<td align="center">54.0</td>
<td align="center">49.9</td>
</tr>

<tr>
<td align="center">ZebraLogic</td>
<td align="center">Acc</td>
<td align="center"><strong>89.0</strong></td>
<td align="center">84.0</td>
<td align="center">37.7<sup>*</sup></td>
<td align="center">73.7</td>
<td align="center">59.3</td>
<td align="center">58.5</td>
<td align="center">57.9</td>
</tr>

<tr>
<td align="center">AutoLogi</td>
<td align="center">Acc</td>
<td align="center"><ins><strong>89.5</strong></ins></td>
<td align="center">88.9</td>
<td align="center">83.3</td>
<td align="center"><strong>89.8</strong></td>
<td align="center">86.1</td>
<td align="center">88.2</td>
<td align="center">84.1</td>
</tr>

<tr>
<td align="center">GPQA-Diamond</td>
<td align="center">Avg@8</td>
<td align="center"><strong>75.1</strong></td>
<td align="center">68.4<sup>*</sup></td>
<td align="center">62.9<sup>*</sup></td>
<td align="center">70.0<sup>*</sup></td>
<td align="center">74.9<sup>*</sup></td>
<td align="center">66.3</td>
<td align="center">68.2</td>
</tr>

<tr>
<td align="center">SuperGPQA</td>
<td align="center">Acc</td>
<td align="center"><strong>57.2</strong></td>
<td align="center">53.7</td>
<td align="center">50.2</td>
<td align="center">55.7</td>
<td align="center">56.5</td>
<td align="center">50.8</td>
<td align="center">49.6</td>
</tr>

<tr>
<td align="center">Humanity's Last Exam<br><sup>(Text Only)</sup></td>
<td align="center">-</td>
<td align="center">4.7</td>
<td align="center">5.2</td>
<td align="center"><ins><strong>5.7</strong></ins></td>
<td align="center">5.8</td>
<td align="center"><strong>7.1</strong></td>
<td align="center">3.7</td>
<td align="center">5.6</td>
</tr>

<tr>
<td align="center" colspan=9><strong>General Tasks</strong></td>
</tr>

<tr>
<td align="center">MMLU</td>
<td align="center">EM</td>
<td align="center"><ins><strong>89.5</strong></ins></td>
<td align="center">89.4</td>
<td align="center">87.0</td>
<td align="center">91.5</td>
<td align="center"><strong>92.9</strong></td>
<td align="center">90.4</td>
<td align="center">90.1</td>
</tr>

<tr>
<td align="center">MMLU-Redux</td>
<td align="center">EM</td>
<td align="center"><ins><strong>92.7</strong></ins></td>
<td align="center">90.5</td>
<td align="center">89.2</td>
<td align="center">93.6</td>
<td align="center"><strong>94.2</strong></td>
<td align="center">92.4</td>
<td align="center">90.6</td>
</tr>

<tr>
<td align="center">MMLU-Pro</td>
<td align="center">EM</td>
<td align="center">81.1</td>
<td align="center"><ins><strong>81.2</strong></ins><sup>*</sup></td>
<td align="center">77.3</td>
<td align="center">83.7</td>
<td align="center"><strong>86.6</strong></td>
<td align="center">81.8</td>
<td align="center">79.4</td>
</tr>

<tr>
<td align="center">IFEval</td>
<td align="center">Prompt Strict</td>
<td align="center"><strong>89.8</strong></td>
<td align="center">81.1</td>
<td align="center">83.2<sup>*</sup></td>
<td align="center">87.6</td>
<td align="center">87.4</td>
<td align="center">88.0</td>
<td align="center">84.3</td>
</tr>

<tr>
<td align="center">Multi-Challenge</td>
<td align="center">Acc</td>
<td align="center"><strong>54.1</strong></td>
<td align="center">31.4</td>
<td align="center">34.0</td>
<td align="center">46.8</td>
<td align="center">49.0</td>
<td align="center">36.4</td>
<td align="center">39.5</td>
</tr>

<tr>
<td align="center">SimpleQA</td>
<td align="center">Correct</td>
<td align="center"><ins><strong>31.0</strong></ins></td>
<td align="center">27.7</td>
<td align="center">13.2</td>
<td align="center">15.9</td>
<td align="center">22.8</td>
<td align="center"><strong>42.3</strong></td>
<td align="center">23.3</td>
</tr>

<tr>
<td align="center">Livebench</td>
<td align="center">Pass@1</td>
<td align="center"><strong>76.4</strong></td>
<td align="center">72.4</td>
<td align="center">67.6</td>
<td align="center">74.8</td>
<td align="center">74.6</td>
<td align="center">69.8</td>
<td align="center">67.8</td>
</tr>
</tbody>
</table>
</div>
<sup>
• Bold denotes global SOTA, and underlined denotes open-source SOTA.
</sup><br/><sup>
• Data points marked with * are taken directly from the model's tech report or blog.
</sup><br/><sup>
• All metrics, except for SWE-bench Verified (Agentless), are evaluated with an 8k output token length. SWE-bench Verified (Agentless) is limited to a 16k output token length.
</sup><br/><sup>
• Kimi K2 achieves 65.8% pass@1 on the SWE-bench Verified tests with bash/editor tools (single-attempt patches, no test-time compute). It also achieves a 47.3% pass@1 on the SWE-bench Multilingual tests under the same conditions. Additionally, we report results on SWE-bench Verified tests (71.6%) that leverage parallel test-time compute by sampling multiple sequences and selecting the single best via an internal scoring model.
</sup><br/><sup>
• To ensure the stability of the evaluation, we employed avg@k on the AIME, HMMT, CNMO, PolyMath-en, GPQA-Diamond, EvalPlus, Tau2.
</sup><br/><sup>
• Some data points have been omitted due to prohibitively expensive evaluation costs.
    </sup>

---

#### Base model evaluation results

<div align="center">

<table>
<thead>
<tr>
<th align="center">Benchmark</th>
<th align="center">Metric</th>
<th align="center">Shot</th>
<th align="center">Kimi K2 Base</th>
<th align="center">Deepseek-V3-Base</th>
<th align="center">Qwen2.5-72B</th>
<th align="center">Llama 4 Maverick</th>
</tr>
</thead>
<tbody>
<tr>
<td align="center" colspan="7"><strong>General Tasks</strong></td>
</tr>
<tr>
<td align="center">MMLU</td>
<td align="center">EM</td>
<td align="center">5-shot</td>
<td align="center"><strong>87.8</strong></td>
<td align="center">87.1</td>
<td align="center">86.1</td>
<td align="center">84.9</td>
</tr>
<tr>
<td align="center">MMLU-pro</td>
<td align="center">EM</td>
<td align="center">5-shot</td>
<td align="center"><strong>69.2</strong></td>
<td align="center">60.6</td>
<td align="center">62.8</td>
<td align="center">63.5</td>
</tr>
<tr>
<td align="center">MMLU-redux-2.0</td>
<td align="center">EM</td>
<td align="center">5-shot</td>
<td align="center"><strong>90.2</strong></td>
<td align="center">89.5</td>
<td align="center">87.8</td>
<td align="center">88.2</td>
</tr>
<tr>
<td align="center">SimpleQA</td>
<td align="center">Correct</td>
<td align="center">5-shot</td>
<td align="center"><strong>35.3</strong></td>
<td align="center">26.5</td>
<td align="center">10.3</td>
<td align="center">23.7</td>
</tr>
<tr>
<td align="center">TriviaQA</td>
<td align="center">EM</td>
<td align="center">5-shot</td>
<td align="center"><strong>85.1</strong></td>
<td align="center">84.1</td>
<td align="center">76.0</td>
<td align="center">79.3</td>
</tr>
<tr>
<td align="center">GPQA-Diamond</td>
<td align="center">Avg@8</td>
<td align="center">5-shot</td>
<td align="center">48.1</td>
<td align="center"><strong>50.5</strong></td>
<td align="center">40.8</td>
<td align="center">49.4</td>
</tr>
<tr>
<td align="center">SuperGPQA</td>
<td align="center">EM</td>
<td align="center">5-shot</td>
<td align="center"><strong>44.7</strong></td>
<td align="center">39.2</td>
<td align="center">34.2</td>
<td align="center">38.8</td>
</tr>
<tr>
<td align="center" colspan="7"><strong>Coding Tasks</strong></td>
</tr>
<tr>
<td align="center">LiveCodeBench v6</td>
<td align="center">Pass@1</td>
<td align="center">1-shot</td>
<td align="center"><strong>26.3</strong></td>
<td align="center">22.9</td>
<td align="center">21.1</td>
<td align="center">25.1</td>
</tr>
<tr>
<td align="center">EvalPlus</td>
<td align="center">Pass@1</td>
<td align="center">-</td>
<td align="center"><strong>80.3</strong></td>
<td align="center">65.6</td>
<td align="center">66.0</td>
<td align="center">65.5</td>
</tr>
<tr>
<td align="center" colspan="7"><strong>Mathematics Tasks</strong></td>
</tr>
<tr>
<td align="center">MATH</td>
<td align="center">EM</td>
<td align="center">4-shot</td>
<td align="center"><strong>70.2</strong></td>
<td align="center">60.1</td>
<td align="center">61.0</td>
<td align="center">63.0</td>
</tr>
<tr>
<td align="center">GSM8k</td>
<td align="center">EM</td>
<td align="center">8-shot</td>
<td align="center"><strong>92.1</strong></td>
<td align="center">91.7</td>
<td align="center">90.4</td>
<td align="center">86.3</td>
</tr>
<tr>
<td align="center" colspan="7"><strong>Chinese Tasks</strong></td>
</tr>
<tr>
<td align="center">C-Eval</td>
<td align="center">EM</td>
<td align="center">5-shot</td>
<td align="center"><strong>92.5</strong></td>
<td align="center">90.0</td>
<td align="center">90.9</td>
<td align="center">80.9</td>
</tr>
<tr>
<td align="center">CSimpleQA</td>
<td align="center">Correct</td>
<td align="center">5-shot</td>
<td align="center"><strong>77.6</strong></td>
<td align="center">72.1</td>
<td align="center">50.5</td>
<td align="center">53.5</td>
</tr>
</tbody>
</table>
</div>
<sup>
• We only evaluate open-source pretrained models in this work. We report results for Qwen2.5-72B because the base checkpoint for Qwen3-235B-A22B was not open-sourced at the time of our study.
</sup><br/><sup>
• All models are evaluated using the same evaluation protocol.

</sup>


## 4. Deployment
> [!Note]
> You can access Kimi K2's API on https://platform.moonshot.ai , we provide OpenAI/Anthropic-compatible API for you.
>
> The Anthropic-compatible API maps temperature by `real_temperature = request_temperature * 0.6` for better compatible with existing applications.

Our model checkpoints are stored in the block-fp8 format, you can find it on [Huggingface](https://huggingface.co/moonshotai/Kimi-K2-Instruct).

Currently, Kimi-K2 is recommended to run on the following inference engines:

* vLLM
* SGLang
* KTransformers
* TensorRT-LLM

Deployment examples for vLLM and SGLang can be found in the [Model Deployment Guide](docs/deploy_guidance.md).

---

## 5. Model Usage

### Chat Completion

Once the local inference service is up, you can interact with it through the chat endpoint:

```python
def simple_chat(client: OpenAI, model_name: str):
    messages = [
        {"role": "system", "content": "You are Kimi, an AI assistant created by Moonshot AI."},
        {"role": "user", "content": [{"type": "text", "text": "Please give a brief self-introduction."}]},
    ]
    response = client.chat.completions.create(
        model=model_name,
        messages=messages,
        stream=False,
        temperature=0.6,
        max_tokens=256
    )
    print(response.choices[0].message.content)
```

> [!NOTE]
> The recommended temperature for Kimi-K2-Instruct is `temperature = 0.6`.
> If no special instructions are required, the system prompt above is a good default.

---

### Tool Calling

Kimi-K2-Instruct has strong tool-calling capabilities.
To enable them, you need to pass the list of available tools in each request, then the model will autonomously decide when and how to invoke them.

The following example demonstrates calling a weather tool end-to-end:

```python
# Your tool implementation
def get_weather(city: str) -> dict:
    return {"weather": "Sunny"}

# Tool schema definition
tools = [{
    "type": "function",
    "function": {
        "name": "get_weather",
        "description": "Retrieve current weather information. Call this when the user asks about the weather.",
        "parameters": {
            "type": "object",
            "required": ["city"],
            "properties": {
                "city": {
                    "type": "string",
                    "description": "Name of the city"
                }
            }
        }
    }
}]

# Map tool names to their implementations
tool_map = {
    "get_weather": get_weather
}

def tool_call_with_client(client: OpenAI, model_name: str):
    messages = [
        {"role": "system", "content": "You are Kimi, an AI assistant created by Moonshot AI."},
        {"role": "user", "content": "What's the weather like in Beijing today? Use the tool to check."}
    ]
    finish_reason = None
    while finish_reason is None or finish_reason == "tool_calls":
        completion = client.chat.completions.create(
            model=model_name,
            messages=messages,
            temperature=0.6,
            tools=tools,          # tool list defined above
            tool_choice="auto"
        )
        choice = completion.choices[0]
        finish_reason = choice.finish_reason
        if finish_reason == "tool_calls":
            messages.append(choice.message)
            for tool_call in choice.message.tool_calls:
                tool_call_name = tool_call.function.name
                tool_call_arguments = json.loads(tool_call.function.arguments)
                tool_function = tool_map[tool_call_name]
                tool_result = tool_function(**tool_call_arguments)
                print("tool_result:", tool_result)

                messages.append({
                    "role": "tool",
                    "tool_call_id": tool_call.id,
                    "name": tool_call_name,
                    "content": json.dumps(tool_result)
                })
    print("-" * 100)
    print(choice.message.content)
```

The `tool_call_with_client` function implements the pipeline from user query to tool execution.
This pipeline requires the inference engine to support Kimi-K2’s native tool-parsing logic.
For streaming output and manual tool-parsing, see the [Tool Calling Guide](docs/tool_call_guidance.md).

---

## 6. License

Both the code repository and the model weights are released under the [Modified MIT License](LICENSE).

---

## 7. Third Party Notices

See [THIRD PARTY NOTICES](THIRD_PARTY_NOTICES.md)

---

## 7. Contact Us

If you have any questions, please reach out at [[email protected]](mailto:[email protected]).