Share your thoughts, 1 month free Claude Pro on usSee more

Instruction Following on IFEval (Accuracy)

90.39Accuracy (IFEval)

Self Consistency (Best on Validation)

Updated 2d ago

Evaluation Results

Method
Self Consistency (Best on Validation) 2025.07	90.39	-
Llama-3.3-70B-Instruct 2025.07	90	-
Simple Router 2025.07	90	-
Open-source Upper Bound 2025.07	90	-
SMCS 2025.07	90	-
MoA 2025.07	89.33	-
Symbolic-MoE* 2025.07	89	-
Claude-3.7-Sonnet 2025.07	88	-
Qwen-2.5-72B-Instruct 2025.07	86.3	-
Qwen3-30B-A3B 2026.05	86.3	-
GPT-4.1 2025.07	86	-
Self-MoA 2025.07	86	-
ZEDASFT 2026.05	85.2	-
Qwen2.5-Instruct 2025.05	85.01	-
ZEDA 2026.05	84.3	-
Qwen3-32B 2025.07	83.7	-
GLM-Z1-32B-0414 2025.07	83	-
Llama-3.3-Nemotron-Super-49B-v1 2025.07	82.7	-
AdaMoE 2026.05	82.4	-
NETSFT 2026.05	82.4	-
GPT-4o 2025.07	82.3	-
Mistral-Small 2025.05	82.25	-
GPT-o3-mini 2025.07	82	-
TeleChat2-35B-32K 2025.07	82	-
QwQ-32B 2025.07	81.7	-
NETSFT→OPD 2026.05	81.7	-
Gemma-3-27b-it 2025.07	81	-
AMARIS 2026.05	81	-
DashAttention 2026.05	80.7	-
Majority Voting 2025.07	80.67	-
AMARIS 2026.05	80.6	-
Rubric-ARM 2026.05	80.4	-
Claude-3.5-Sonnet 2025.07	80.3	-
DeepSeek-R1-Distill-Llama-70B 2025.07	80.3	-
Qwen2.5-Coder-32B-Instruct 2025.07	80.3	-
InfiGFusion 2025.05	80.22	-
FullAttn 2026.05	80	-
NSA 2026.05	80	-
RubricHub (RuRL) 2026.05	79.8	-
OpenRubrics 2026.05	79.5	-
MiniLogit 2025.05	79.36	-
RuscaRL 2026.05	79	-
InfLLMv2 2026.05	79	-
FuseChat 2025.05	78.9	-
Qwen2.5-32b-Instruct 2025.07	78.7	-
RLCF 2026.05	78.6	-
FuseLLM 2025.05	78.56	-
Pivot-SFT 2025.05	77.7	-
Phi-4 2025.05	77.34	-
Naive (no RL) 2026.05	77.3	-
EXAONE-Deep-32B 2025.07	76.3	-
InfiFusion 2025.05	76.02	-
Qwen2.5-Coder 2025.05	74.7	-
HuatuoGPT-o1-72B 2025.07	74	-
DeepSeek-R1-Distill-Qwen-32B 2025.07	73.7	-
Dynamic Skipping 2026.05	70.4	-
ZEDA 2026.05	68.2	-
GLM-4.7-Flash 2026.05	67.3	-
ZEDASFT 2026.05	66.4	-
NETSFT 2026.05	65.8	-
NETSFT→OPD 2026.05	65.1	-
InternLM2.5-20B-Chat 2025.07	64.7	-
AdaMoE 2026.05	63.8	-
Dynamic Skipping 2026.05	63.4	-
Baseline 2026.05	58.41	-
Elastic-dLLM 2026.05	58.04	-
IBTPO 2026.05	57.9	-
DPad 2026.05	57.86	-
SparseD 2026.05	57.67	-
IBRO 2026.05	54.3	-
Vanilla GRPO 2026.05	53.9	-
TreeRL 2026.05	53.4	-
Initial Model 2026.05	52.9	-
Baseline 2026.05	52.68	-
GRPO w/ Entropy Reg 2026.05	52.5	-
SparseD 2026.05	52.31	-
Elastic-dLLM 2026.05	52.31	-
DPad 2026.05	51.2	-
IBTPO 2026.05	46.6	-
IBTPO 2026.05	46.4	-
Initial Model 2026.05	43.5	-
FP16 2025.11	43.44	-
Initial Model 2026.05	43.3	-
TreeRL 2026.05	42.6	-
Vanilla GRPO 2026.05	42.4	-
Vanilla GRPO 2026.05	42.3	-
TreeRL 2026.05	42.2	-
Quant-Only 2025.11	41.4	-
IntAttention 2025.11	39.74	-
Gemma 3 270M 2026.05	-	31.2
SmolLM2 360M 2026.05	-	40.2
MobileMoE-S 2026.05	-	54.1
Qwen3.5 0.8B 2026.05	-	59.9
MobileMoE-M 2026.05	-	60.8
Gemma 3 1B 2026.05	-	63.7
MobileLLM-Pro 2026.05	-	63.1
Llama 3.2 1B 2026.05	-	58.7
OLMo 2 1B 2026.05	-	53.5
SmolLM2 1.7B 2026.05	-	54.7
Qwen3.5 2B 2026.05	-	73.3

Showing 100 of 102 rows