DataLearner logo

DeepSeek-V4-ProvsKimi K2.6

Across 9 shared benchmarks, Kimi K2.6 leads overall: DeepSeek-V4-Pro wins 1, Kimi K2.6 wins 8, with 0 ties and an average score difference of -25.21.

DeepSeek-AI
DeepSeek-V4-Pro

DeepSeek-AI · 2026-08-13 · Reasoning model

Moonshot AI
Kimi K2.6

Moonshot AI · 2026-04-20 · Reasoning model

DeepSeek-V4-Pro1 win(11%)(89%)8 winsKimi K2.6

Benchmark scores

Grouped by capability, sorted by largest gap within each. 9 shared benchmarks.

Agent Level Benchmark

Kimi K2.6 2/2
BenchmarkDeepSeek-V4-ProKimi K2.6Diff
τ²-Bench - Telecom91.2059 / 264Normal (With Tools)93.9043 / 264Normal (With Tools)-2.70
Terminal Bench Hard36.4067 / 244Normal (With Tools)37.9059 / 244Normal (With Tools)-1.50

General Knowledge

Kimi K2.6 2/2
BenchmarkDeepSeek-V4-ProKimi K2.6Diff
HLE8.20416 / 563Normal (No Tools) · Text only19.60290 / 563Normal (No Tools) · Text only-11.40
CritPt0.90158 / 200Normal (No Tools)1.40135 / 200Normal (No Tools)-0.50

General Evaluation

Kimi K2.6 1/1
BenchmarkDeepSeek-V4-ProKimi K2.6Diff
GPQA Diamond72.90301 / 462Normal (No Tools)78.80249 / 462Normal (No Tools)-5.90

Instruction Following

DeepSeek-V4-Pro 1/1
BenchmarkDeepSeek-V4-ProKimi K2.6Diff
IF Bench45.80176 / 282Normal (No Tools)44.30184 / 282Normal (No Tools)+1.50

Long Context

Kimi K2.6 1/1
BenchmarkDeepSeek-V4-ProKimi K2.6Diff
AA-LCR53135 / 170Normal (No Tools)69.70106 / 170Normal (No Tools)-16.70

Text Embedding

Kimi K2.6 1/1
BenchmarkDeepSeek-V4-ProKimi K2.6Diff
Context Arena31.43115 / 126Normal (No Tools)51.8886 / 126Normal (No Tools)-20.45

Writing and Creative Capabilities

Kimi K2.6 1/1
BenchmarkDeepSeek-V4-ProKimi K2.6Diff
Creative Writing1,55247 / 106Normal (No Tools)1,72125 / 106Normal (No Tools)-169.20

Specs

FieldDeepSeek-V4-ProKimi K2.6
PublisherDeepSeek-AIMoonshot AI
Release date2026-08-132026-04-20
Model typeReasoning modelReasoning model
ArchitectureMoEMoE
Parameters1.6T1T
Context length1M256K
Max output384KNot available

API pricing

Prices use DataLearner records when available; missing fields are not inferred.

ItemDeepSeek-V4-ProKimi K2.6
Text input$0.435 / 1M tokens$0.95 / 1M tokens
Text output$0.87 / 1M tokens$4 / 1M tokens
Cache read$0.003625 / 1M tokens$0.16 / 1M tokens
Cache writeNot public$0.95 / 1M tokens

Summary

  • DeepSeek-V4-Proleads in:Instruction Following (1/1)
  • Kimi K2.6leads in:Agent Level Benchmark (2/2), General Knowledge (2/2), General Evaluation (1/1), Long Context (1/1), Text Embedding (1/1), Writing and Creative Capabilities (1/1)

On average across the 9 shared benchmarks, Kimi K2.6 scores 25.21 higher.

Largest single-benchmark gap: Creative Writing — DeepSeek-V4-Pro 1,552 vs Kimi K2.6 1,721 (-169.20).

Page generated from structured model, pricing and benchmark records. No real-time LLM is used to write the prose.