Head to head

Any two models from the board, on one scale.

Model A

Model B

openrouter · qwenhosted

overall

6 ▸

openrouter · deepseekhosted

54vs60

DeepSeek V4.1 Flash leads by 6 overall · ahead on 7 of 9 tasks

tasks

Task by task

9 tasks

79TOOL◂ +178

47STRUCT+5 ▸52

41RAG+22 ▸63

89CTX+11 ▸100

75CODE◂ +3441

32REASON+7 ▸39

25INSTR+21 ▸46

38CLASS+17 ▸55

60SUM+10 ▸70

best build per side · same suite, same graders · hover a score for cases passed

specs

64 tok/sspeed215 tok/s

—memory—

1Mcontext1.0M

$0.5705cost / run$0.7017

—quantization—

openrouterharnessopenrouter

the lit value is the better spec — faster, smaller, cheaper, or more context · quantization & harness just describe the builds