Which is better, Kimi K2.6 or Qwen3.5 4B?

Kimi K2.6 vs Qwen3.5 4B

26 SHARED BENCHMARKS

Across 26 shared benchmarks, Kimi K2.6 scores higher on 25 and Qwen3.5 4B on 1. The widest gap is gdpval, where Kimi K2.6 scores 50.4 against 8.3. Qwen3.5 4B is the cheaper of the two on tracked API pricing ($0.03 against $0.95 per million input tokens).

MOONSHOTVSALIBABA26 SHARED25–1 HEAD-TO-HEAD

BenchmarkKimi K2.6MARGINQwen3.5 4B

AA Agentic Index58.736.3

AA Intelligence44.220

AA-LCR8163

AA-Omniscience6.4-58.4

aime_202696.499

Artificial Analysis Coding Index61.822.6

critpt80

gdpval50.48.3

GPQA Diamond91.177.1

HLE37.59.9

hmmt_feb_202694.784.1

IFBench7659.2

LiveCodeBench v690.255.8

MMLU-Pro88.179.1

MMMU-Pro80.165.4

multichallenge63.149

OmniScience Accuracy32.815.1

OmniScience Non-Hallucination60.713.4

PinchBench90.271.3

scicode53.518.3

TauBench V3 - Banking23.36.8

Terminal-Bench 2.167.225.8

Terminal-Bench Hard43.918.2

τ²-Bench Telecom (AA run)95.992.1

τ³-Bench23.18.2

τ³-Bench Banking20.65.5

Best tracked score per model per benchmark (default configuration; source-attributed). ↓ marks lower-is-better metrics. Open either model for its full surface, provenance and pricing. Quoted rates are the price-setter row we currently track for each model — its direct or vendor-official listing where one exists (direct, direct), otherwise the lowest tracked offer.

Compare Kimi K2.6 withALL PAIRINGS →

MoonshotKimi K2.5 Kimi K3 Kimi K2 Instruct

Across the fieldGemini 3.6 Flash Nemotron 3 Super 120B A12B GPT-5.5 Claude Opus 4.8 GPT-4o GPT-5.6 Sol Gemini 3.1 Pro GPT-5.4

Compare Qwen3.5 4B withALL PAIRINGS →

AlibabaQwen3.5 35B A3B Qwen3 VL 4B (Reasoning)Qwen3.5 397B A17B Qwen2.5 Instruct 72B Qwen3.5 27B Qwen3.6 Plus

Across the fieldKimi K2.5 o3 GPT-5.1 Claude Opus 5 Gemma 4 26B A4B Kimi K3 Llama 3.1 Instruct 405B DeepSeek-V3.2