deepseek-v3.2 vs kimi-k2
Side-by-side comparison of deepseek-v3.2 and kimi-k2: benchmarks, pricing, context window and capabilities. Both are accessible through Requesty's unified API. deepseek-v3.2 outperforms kimi-k2 on 5 of 9 shared benchmarks.
deepseek-v3.2
Input / 1M
$0.56
Output / 1M
$1.68
Context
164K
Model ID
vertex/deepseek-v3.2
kimi-k2
Input / 1M
$0.60
Output / 1M
$2.50
Context
262K
Model ID
vertex/kimi-k2
Benchmark comparison
Intelligence Indexreasoning
deepseek-v3.225.3%
kimi-k225.9%
Coding Indexcoding
deepseek-v3.244.2%
kimi-k2N/A
Math Indexmath
deepseek-v3.292.0%
kimi-k294.7%
GPQA Diamondreasoning
deepseek-v3.284.0%
kimi-k283.8%
AIME 2025math
deepseek-v3.292.0%
kimi-k294.7%
LiveCodeBenchcoding
deepseek-v3.286.2%
kimi-k285.3%
Terminal-Bench Hardagentic
deepseek-v3.235.6%
kimi-k231.1%
τ²-Benchagentic
deepseek-v3.290.6%
kimi-k293.0%
MMLU Proknowledge
deepseek-v3.286.2%
kimi-k284.8%
Humanity's Last Examreasoning
deepseek-v3.224.6%
kimi-k223.8%
Scores sourced from official model cards, Artificial Analysis, and public leaderboards. Benchmarks measure specific skills and don't capture every aspect of model quality.
Pricing & specifications
| deepseek-v3.2 | kimi-k2 | |
|---|---|---|
| Input price / 1M | $0.56 | $0.60 |
| Output price / 1M | $1.68 | $2.50 |
| Context window | 164K tokens | 262K tokens |
| Max output | 66K tokens | 262K tokens |
| Vision input | N/A | N/A |
| Tool calling | Yes | Yes |
| Reasoning | Yes | Yes |
| Prompt caching | Yes | Yes |
| Computer use | N/A | N/A |
| Provider | Google LLC (Vertex AI) | Google LLC (Vertex AI) |
Questions people ask
Is deepseek-v3.2 better than kimi-k2?
deepseek-v3.2 outperforms kimi-k2 on 5 of 9 shared benchmarks. See the benchmark comparison above for specifics: deepseek-v3.2 and kimi-k2 have different strengths across reasoning, coding, math and multimodal tasks.
Which is cheaper, deepseek-v3.2 or kimi-k2?
deepseek-v3.2 is cheaper. deepseek-v3.2 costs $0.56/$1.68 per 1M input/output tokens, while kimi-k2 costs $0.60/$2.50.
Can I use deepseek-v3.2 and kimi-k2 through the same API?
Yes. Requesty provides a single OpenAI-compatible API that routes to both. Change just the "model" parameter to switch between "vertex/deepseek-v3.2" and "vertex/kimi-k2", no other code changes needed.
What are the context windows?
deepseek-v3.2 supports up to 164K tokens of context. kimi-k2 supports up to 262K tokens. Longer context means you can feed larger documents or codebases in a single prompt, though quality often degrades past 128K for most models.
Switch between deepseek-v3.2 and kimi-k2 with one line of code
Requesty provides a single OpenAI-compatible API for 600+ models. Change the model parameter, not your code.