Which model leads BFCLv4?

Q: Which model leads BFCLv4?

Llama 3.1 Instruct 405B tops the board at 88.5

BFCLv4 Leaderboard — 43 Models Ranked

Pro Mode — raw scores, one benchmark under the lensofficial card harness

The field · 43 measured

7.1788.5

43 models measured, most on the official card harness. Llama 3.1 Instruct 405B tops the board at 88.5.

43 measured·Lifecycle Updated Oct 9, 2026Score basis

#◆Vendor$/M in$/M out

Leader — ignites whiteField — fill cools with scoreIndependently verified only0 of 43Bars show 7.17–88.5 · ◆ solid = first-party or better · outlined = arm's-length

01Llama 3.1 Instruct 405BMeta88.5———

02Llama 3.1 70B InstructMeta84.8$0.56/M$0.56/M$0.007

03Atria Dawnshanghai-ai-lab77———

04Llama 3.1 8B InstructMeta76.1$0.02/M$0.04/M$0.000

05Qwen3.7 MaxAlibaba75$2.5/M$7.5/M$0.067

06Ling-3.0-flashInclusionAI73$0.07/M$0.22/M$0.002

07Qwen3.7 Plus PreviewAlibaba72.9$0.40/M$1.6/M$0.014

08Qwen3.5 397B A17BAlibaba72.9$0.60/M$3.6/M$0.029

09Qwen3.5 122B A10BAlibaba72.2$0.40/M$3.2/M$0.025

10Qwen3 235B A22BAlibaba70.8$0.70/M$2.8/M$0.025

11Qwen3 32BAlibaba70.3$0.16/M$0.64/M$0.006

12Qwen3 30B A3BAlibaba69.1$0.20/M$0.80/M$0.007

13Qwen3.5 27BAlibaba68.5$0.30/M$2.4/M$0.020

14Nova Pro+1 altAmazon68.4$0.80/M$3.2/M$0.029

15Qwen3 Max (Reasoning)Alibaba67.7$1.2/M$6.0/M$0.053

16Qwen3.5 35B A3BAlibaba67.3$0.25/M$2.0/M$0.017

17Nova Lite+1 altAmazon66.6$0.06/M$0.24/M$0.002

18QwQ 32B PreviewAlibaba66.4$0.66/M$1.0/M$0.012

19Granite 4.2 30B+1 altIBM61.39$0.16/M$0.65/M$0.007

20Qwen3.5 9B+2 altsAlibaba60.13$0.14/M$0.20/M$0.003

21Seed 1.8ByteDance57.2$0.25/M$2.0/M$0.020

22LFM2.5-2.6B+2 altsLiquid AI56.88———

23Nova Micro+1 altAmazon56.2$0.04/M$0.14/M$0.002

24Granite 4.2 3B+2 altsIBM52.41$0.03/M$0.12/M$0.001

25Qwen3.5 4B+4 altsAlibaba50.56$0.03/M$0.15/M$0.002

26Qwen3 30B A3B 2507 ThinkingAlibaba50.533P$0.20/M$2.4/M$0.026

27Granite 4.2 8B+2 altsIBM50.29$0.06/M$0.25/M$0.003

28LFM2.5-8B-A1BLiquid AI49.733P———

29Gemma 4 E4B+3 altsGoogle46.39$0.02/M$0.10/M$0.001

30Qwen3.5 2B+1 altAlibaba43.6———

31Gemma 4 E2B+4 altsGoogle36.98———

32LFM2.5-VL-3B+2 altsLiquid AI32.5———

33Granite 4.0 H TinyIBM28.523P———

34LFM2-8B-A1B (Non-Reasoning)Liquid AI25.523P———

35Qwen3.5 0.8B+1 altAlibaba25.3———

36LFM2.5-350M+1 altLiquid AI21.863P———

37LFM2.5-VL-450M-ExtractLiquid AI21.083P———

38LFM2.5-230MLiquid AI21.033P———

39LFM2-VL-3B+1 altLiquid AI20.5———

40Qwen3.5 0.8B (Instruct)+1 altAlibaba18.73P———

41Granite 4.0 350M+1 altIBM13.733P———

42LFM2-350M+1 altLiquid AI12.293P———

43Gemma 3 1B Instruct+1 altGoogle7.173P———

◆ 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: 43 models scored · 0 independently verified · 7 vendor cross-reference · 36 vendor-reported · 3 with source disagreement. How these tiers are assigned