Compare AI models
| Specifications | ||
ParametersThe learned values that shape how a model responds. B means billion and T means trillion. For mixture-of-experts models, the total count is larger than the number active for each token. Parameter count alone does not measure quality. | 1T | 4B/30B/235B |
Context windowHow much information the model can work with in one request, measured in tokens — pieces of words or other data. This includes your question, attachments, conversation and reply. Limits can vary by model variant and provider. | 256k | 256k |
| API pricingUSD per 1M tokens · lower wins | ||
Cheapest inputLowest input rate across every provider on OpenRouter, the lab included. The cheapest endpoint may run a quantised build or a shorter context — see "Available from" on the model page. | $0.60Novita | — |
Cheapest outputLowest output rate across every provider on OpenRouter, the lab included. May come from a different provider than the cheapest input. | $2.50Novita | — |
| $245Novita · bf16 | — | |
| Benchmarks | ||
LiveCodeBenchCompetitive coding — Coding problems published so recently the AI can't have seen them in training — a contamination-free test of raw programming skill. Higher is better. | 83.1% | 74.1% |
Humanity's Last Exam · text only, no toolsMultidisciplinary reasoning — accuracy on the 2,158 text-only questions without tools | 23.9% | 18.2% |
AIME 2025Reasoning — single-answer accuracy on the 2025 invitational mathematics exam without tools | 94.5% | 92.3% |
GPQA DiamondScience — Graduate-level science questions in biology, physics, and chemistry — hard enough that subject-matter PhDs score around 65%. Higher is better. | 84.5% | 81.1% |
MMLU-ProKnowledge — accuracy on the ten-option MMLU-Pro benchmark | 84.6% | 84.4% |
SWE-Bench VerifiedCoding — Real coding tasks pulled from open-source projects — the AI has to find and fix actual bugs. A human-checked version of the original SWE-Bench. Higher is better. | 71.3% | — |
SWE-Bench MultilingualMultilingual coding — Like SWE-Bench, but the coding problems span many programming languages, not just one. Tests how broadly the AI can code. Higher is better. | 61.1% | — |
IFEvalAgentic tasks — following verifiable instructions in generated responses | — | 87.8% |
BrowseCompWeb browsing — Can the AI browse the web and track down hard-to-find answers? Higher is better. | 60.2% | — |
| Overview | ||
| Company | Moonshot AI | Qwen |
| Release date | Nov 6 2025 | Jul 21 2025 |
| Access | Open Weight | Open Weight |
| Model details | View model | View model |
Other comparisons
Kimi K2 ThinkingvsClaude Fable 5.2Qwen3-2507vsClaude Fable 5.2Kimi K2 ThinkingvsGPT-6.2 AstraQwen3-2507vsGPT-6.2 AstraKimi K2 ThinkingvsGemini 4 NeonQwen3-2507vsGemini 4 NeonKimi K2 ThinkingvsMuse EmberQwen3-2507vsMuse EmberKimi K2 ThinkingvsGrok 4.7Qwen3-2507vsGrok 4.7Kimi K2 ThinkingvsDeepSeek-V4.1-Flash-1017Qwen3-2507vsDeepSeek-V4.1-Flash-1017Frequently asked questions
Kimi K2 Thinking leads Qwen3-2507 on 5 of the 5 benchmarks they both report. Qwen3-2507 shipped 108 days before Kimi K2 Thinking, so benchmark comparisons should account for the intervening progress.
Kimi K2 Thinking has 1T parameters, while Qwen3-2507 has 4B/30B/235B. Context windows are 256k (Kimi K2 Thinking) vs 256k (Qwen3-2507).
On LiveCodeBench, Kimi K2 Thinking leads at 83.1% vs Qwen3-2507 at 74.1%. On Humanity's Last Exam · text only, no tools, Kimi K2 Thinking leads at 23.9% vs Qwen3-2507 at 18.2%. On AIME 2025, Kimi K2 Thinking leads at 94.5% vs Qwen3-2507 at 92.3%. On GPQA Diamond, Kimi K2 Thinking leads at 84.5% vs Qwen3-2507 at 81.1%. On MMLU-Pro, Kimi K2 Thinking leads at 84.6% vs Qwen3-2507 at 84.4%.