VerdictPal · editorial desk · updated 1 Sep 2026VerdictPal

Compare · Two models, head-to-head

Pick two models. See the diff.

Composite scores, benchmark ranks, capability grids, and pricing per million tokens. All sourced from the same data that powers the model atlas.

Tools Models

Kimi K3 vs MiniMax M3

Model A

Model B

MODEL A

Kimi K3

Moonshot · Kimi K3

Composite 58Intel 7514 rows3 sources

Moonshot AI's 2.8T-parameter multimodal reasoning model for long-horizon coding and knowledge work, with a 1M-token context window, public API access, and open weights on Hugging Face.

MODEL B

MiniMax M3

MiniMax · MiniMax

Composite 63Intel 4517 rows8 sources

MiniMax's May 2026 open-weights release. Terminal-Bench 2.0 66.0% — the #10 score globally and the strongest non-frontier-tier result on the public ledger. 1M context, $0.40 / $2.00 per 1M tokens.

Kimi K31stats won

Tied1stats

MiniMax M35stats won

Spec showdown

Capacity, price, freshness.

Kimi K3vsMiniMax M3

Composite scores are within 5 points — 58 for Kimi K3 vs 63 for MiniMax M3. The choice comes down to capability profile and price. Kimi K3 has the higher Vals/AA intelligence score (75 vs 45). MiniMax M3 is significantly cheaper at $0.40/1M input vs $3.0/1M input.

Composite score
Kimi K358MiniMax M363Winner on this row
Intelligence (Vals/AA)
Kimi K375Winner on this rowMiniMax M345
Context window
Kimi K31MMiniMax M31M
Max output
Kimi K3—MiniMax M366KWinner on this row
Input price / 1M (lower wins)
Kimi K3$3.0MiniMax M3$0.40Winner on this row
Output speed (AA default API)
Kimi K340 t/sMiniMax M3114 t/sWinner on this row
Time to first token (AA default API)
Kimi K32.39sMiniMax M30.95sWinner on this row

Composite score

Intelligence (Vals/AA)

Context window

Max output

—

66K

Input price / 1M (lower wins)

$3.0

$0.40

Output speed (AA default API)

40 t/s

114 t/s

Time to first token (AA default API)

2.39s

0.95s

Capability grid

What each model can do.

Modalities

Kimi K3MiniMax M3

Text●●

Image●—

Audio——

Video——

Code●●

Tool use●●

Reasoning●●

Access

Kimi K3MiniMax M3

API●●

Consumer app●—

Open weights●●

Managed cloud—●

On-device——

Family profile

Top score per benchmark family.

Knowledge

Reasoning

Coding

Agentic

Long context

Benchmark ranks

Every shared benchmark, side by side.

Reasoning

Artificial Analysis Intelligence Index#8/184 · #44/184

59.7%

45.4%

GPQA Diamond#8/185 · #13/185

93.5%

92.9%

Humanity's Last Exam#10/180 · #39/180

46.9%

39%

Coding

SciCode#3/180 · #80/180

58.7%

45.4%

Agentic

Terminal-Bench#10/184 · #54/184

85.02%

66%

τ³-Bench Banking#7/106 · #69/106

45.98%

15.26%

Long context

AA-LCR#2/179 · #5/179

82.67%

80.33%

Performance

AA output speed#79/185 · #31/185

40 t/s

114 t/s

AA time to first token#25/185 · #66/185

2.39s

0.95s

Dossier facts

Dates, sources, fineprint.

Kimi K3MiniMax M3

ProviderMoonshotMiniMax

FamilyKimi K3MiniMax

Release date2026-07-162026-05-31

Knowledge cutoffunpublished2026-03-01

Context window1M1M

Max output—66K

Input price /1M$3.0$0.40

Output price /1M$15$2.0

Statussolidsolid

Composite score5863

Intelligence score7545

Intel sourcevalsaa

Benchmark rows1417

Source count38

Freshnessfreshfresh

Pricing checked2026-07-272026-06-01

Ecosystem

Which tools wrap each model.

Kimi K3

Not listed in any tool yet.

MiniMax M3

Not listed in any tool yet.

Verdict

Which model to actually pick.

Open Kimi K3's dossier Open MiniMax M3's dossier