DeepSeek R1 (0528) vs Kimi K2 Thinking

Benchmarks, API pricing and specs, head to head. Data updated 2026-06-10.

DeepSeek R1 (0528)

DeepSeek · May 2025

67.8

The open reasoning model that started it all — RL-trained chain-of-thought, MIT licensed, and a research landmark.

Kimi K2 Thinking

Moonshot AI · Nov 2025

67.6

A trillion-parameter open reasoning agent that can chain 200–300 tool calls — the open-weights agentic standout of late 2025.

The verdict

Kimi K2 Thinking wins 3 of the 4 benchmarks these models share, against 1 for DeepSeek R1 (0528). DeepSeek R1 (0528) is about 1.1x cheaper per blended million tokens (3:1 input:output mix). Kimi K2 Thinking also takes 262K of context versus 128K for DeepSeek R1 (0528).

Benchmark head-to-head 13

GPQA Diamond
81%84.5%
DeepSeek R1 (0528)Kimi K2 Thinking
AIME 2025
87.5%~94.5%
DeepSeek R1 (0528)Kimi K2 Thinking
HLE
17.7%23.9%
DeepSeek R1 (0528)Kimi K2 Thinking
MMLU-Pro
84.8%~84.6%
DeepSeek R1 (0528)Kimi K2 Thinking

Specs & pricing

DeepSeek R1 (0528)Kimi K2 Thinking
modhub Index67.867.6
Input price / 1M$0.55$0.6
Output price / 1M$2.19$2.5
Context window128K262K
Max output64K
Open weightsyes (MIT)yes (Modified MIT)
Reasoning modelyesyes
Multimodal inputtexttext
Knowledge cutoffMar 2025Apr 2025
ReleasedMay 2025Nov 2025
Example monthly cost*$8.79$9.75

* 10M input + 1.5M output tokens per month at list prices, no caching. Green = better value on that row.

Frequently asked questions

Which is better, DeepSeek R1 (0528) or Kimi K2 Thinking?
Kimi K2 Thinking wins 3 of the 4 benchmarks these models share, against 1 for DeepSeek R1 (0528). DeepSeek R1 (0528) is about 1.1x cheaper per blended million tokens (3:1 input:output mix). Kimi K2 Thinking also takes 262K of context versus 128K for DeepSeek R1 (0528).
Which is cheaper, DeepSeek R1 (0528) or Kimi K2 Thinking?
DeepSeek R1 (0528) costs $0.55/$2.19 per million input/output tokens, while Kimi K2 Thinking costs $0.6/$2.5. For a typical workload of 10M input and 1.5M output tokens per month, that's $8.79 versus $9.75.
Which model is better for coding, DeepSeek R1 (0528) or Kimi K2 Thinking?
We don't yet track SWE-bench Verified results for both models; check their individual pages for coding-related scores.

More comparisons