Grok 4.6 vs Kimi K3

At a glance

• Capability index

  • Grok 4.6156.4
  • Kimi K3157.4

  Kimi K3 is 1.0 points higher.

• Price per 1M tokens

  • Grok 4.6$3.00
  • Kimi K3$6.00

  Grok 4.6 is 2 times cheaper.

• Context window

  • Grok 4.6500K tokens
  • Kimi K31M tokens

  Kimi K3 takes in 2.1 times as much.

• Released

  • Grok 4.6Aug 12, 2026
  • Kimi K3Jul 16, 2026

  Grok 4.6 is 27 days newer.

Where each scores higher

On the 22 tests both have taken, each on its own scale. A gap of a point or two can sit within a test’s margin of error.

1 more test has results for only one of them; the table below lists every result.

Everything side by side

Comparison of Grok 4.6 vs Kimi K3

Measure	Grok 4.6xAI	Kimi K3Moonshot AI
Model
Maker	xAI	Moonshot AI
Sold here by	xAI	Moonshot AI
API model ID	grok-4.6	kimi-k3
Released	Aug 12, 2026	Jul 16, 2026
Knowledge cutoff	Feb 1, 2026	Not reported
Weights	Closed: hosted access only	Open: downloadable
Developer’s country	United States	China
Hosts selling it	35including xAI directly	84including Moonshot AI directly
Training compute	Not published	2.0 × 10²⁵ FLOPEpoch AI estimate
Test results
Capability index	156.4#21 of 274	157.4 (best of these models)#15 of 274
LiveBench overall	78.0%#18 of 66	79.2% (best of these models)#15 of 66
Arena text	1454#73 of 413	1488 (best of these models)#16 of 413
LiveBench coding	76.8%#42 of 66	81.5% (best of these models)#15 of 66
Arena WebDev	1620#23 of 138	1658 (best of these models)#13 of 138
Arena coding	1507#64 of 408	1541 (best of these models)#9 of 408
LiveBench mathematics	92.6% (best of these models)#25 of 66	84.4%#53 of 66
Mock AIME 2024–2025	99.2% (best of these models)#15 of 297	97.2%#31 of 297
FrontierMath Tiers 1–3	66.0%#28 of 114	72.2% (best of these models)#22 of 114
FrontierMath Tier 4	31.7%#33 of 70	39.0% (best of these models)#29 of 70
Arena math	1448#78 of 396	1501 (best of these models)#16 of 396
SimpleQA Verified	49.3%#27 of 86	50.6% (best of these models)#24 of 86
LiveBench agentic coding	57.0%#22 of 66	62.2% (best of these models)#10 of 66
DeepSWE	67.5%#20 of 69	68.5% (best of these models)#19 of 69
CursorBench	41.4%#24 of 62	Not tested
LiveBench language	83.7%#19 of 66	85.5% (best of these models)#11 of 66
LiveBench instructions	71.9% (best of these models)#20 of 66	71.4%#23 of 66
Arena creative writing	1445#52 of 411	1460 (best of these models)#26 of 411
Arena instruction following	1451#63 of 413	1488 (best of these models)#14 of 413
LiveBench data analysis	73.9%#40 of 66	78.7% (best of these models)#21 of 66
LiveBench reasoning	90.5%#12 of 66	90.7% (best of these models)#9 of 66
GPQA Diamond	94.0% (best of these models)#11 of 319	93.1%#20 of 319
Chess puzzles	40.0% (best of these models)#24 of 227	39.0%#31 of 227
Arena hard prompts	1479#67 of 413	1517 (best of these models)#9 of 413
Price per million tokens
Input	$2.00 (best of these models)	$3.00
Output	$6.00 (best of these models)	$15.00
Blended, 3 input to 1 output	$3.00 (best of these models)	$6.00
Cached input	$0.50	$0.30 (best of these models)
Long requests	$4.00 in, $12.00 outabove 200K tokens	No separate price reported
The maker’s own price	This listing	This listing
Limits
Context window	500K tokens	1M tokens (best of these models)
Max input	372K tokens	1M tokens (best of these models)
Max output	500K tokens	1M tokens (best of these models)
Features
Reads	images, PDFs and text	images, text and video
Produces	text	text
Reasoning	Yeseffort low, medium, high, xhigh	Yeseffort low, high, max; can be switched off; thinking budget can be set
Tool calling	Yes	Yes
Structured output	Yes, such as JSON	Yes, such as JSON
File attachments	Yes	Yes
Temperature setting	Supported	Fixed by the provider

Bold marks the better value in each row. Prices are each model’s own list price, or the middle price across hosts when the maker does not sell it, blended as three input tokens for every output token. Test results as published by LiveBench (Apache-2.0), Epoch AI (CC BY 4.0), Datacurve DeepSWE, compiled by Epoch AI (CC BY 4.0), Cursor CursorBench, compiled by Epoch AI (CC BY 4.0), Arena (CC BY 4.0); model details and prices from models.dev (MIT). Confirm pricing with the provider before use.