Which model leads BigCodeBench (Hard)?

Q: Which model leads BigCodeBench (Hard)?

Qwen 2.5 Coder 32B Instruct tops the board at 27

BigCodeBench (Hard) Leaderboard — Model Scores

Pro Mode — raw scores, one benchmark under the lensthird-party eval harness

The field · 7 measured

13.527

7 models measured, most on the third-party eval harness. Qwen 2.5 Coder 32B Instruct tops the board at 27.

7 measured·Lifecycle Updated Oct 8, 2026Score basis

#◆Vendor$/M in$/M out

Leader — ignites whiteField — fill cools with scoreIndependently verified only0 of 7Bars show 13.5–27 · ◆ solid = first-party or better · outlined = arm's-length

01Qwen 2.5 Coder 32B InstructAlibaba27VENDOR$0.06/M$0.20/M$0.005

02Seed Coder 8B InstructByteDance26.4———

03Qwen3 8BAlibaba23$0.18/M$0.70/M$0.019

04Qwen2.5 Coder Instruct 7BAlibaba20.3$0.01/M$0.03/M$0.001

05Yi Coder 9B Chat01.AI17.6———

06DeepSeek-Coder-6.7B-InstructDeepSeek15.5———

07Llama 3.1 8B InstructMeta13.5$0.02/M$0.04/M$0.002

◆ solid = independently verified · aggregator attested · vendor attributed — outlined = cross-reference · source-attributed. Dimmed ◆ = confidence inferred. “+N alts” = alternate disclosures in the drawer with why-demoted. Prices per million tokens; $/point = input price ÷ score. Every row opens its model dossier.

Verification: 7 models scored · 0 independently verified · 5 vendor cross-reference · 2 vendor-reported · 0 with source disagreement. How these tiers are assigned