List your company

Multi-choice Question Answering on LongBench v2

46.5Overall AccuracyQwen2.5-14B-1M-LongRLVR
24.430.137535.87541.6125May 19, 2025Jul 5, 2025Aug 22, 2025Oct 9, 2025Nov 26, 2025Jan 13, 2026Mar 2, 2026

Evaluation results

MethodLinks
Qwen2.5-14B-1M-LongRLVRBackbone=Qwen2.5-14B-1M, Protocol=LongRLVR2026.0346.555.643.338
Qwen3-14B (Thinking)Backbone=Qwen3-14B, Thinking Mode=true2026.0344.951.742.338.9
Qwen2.5-72B-YaRNBackbone=Qwen2.5-72B, Extension=YaRN2026.0343.543.548.940.9
QwenLong-L1-32BBackbone=QwenLong-L1-32B2026.034153.334.433.3
Qwen2.5-14B-1MBackbone=Qwen2.5-14B-1M2026.0340.251.73433.3
Qwen2.5-14B-1M-RLVRBackbone=Qwen2.5-14B-1M, Protocol=RLVR2026.0339.848.336.731.5
Qwen2.5-14B-1M-SFTBackbone=Qwen2.5-14B-1M, Protocol=SFT2026.0339.648.934.933.3
Qwen2.5-7B-1M-LongRLVRBackbone=Qwen2.5-7B-1M, Protocol=LongRLVR2026.0338.645.635.832.4
Qwen3-8B (Thinking)Backbone=Qwen3-8B, Thinking Mode=true2026.0337.643.328.832.4
LLaMA-3.1-8B-LongRLVRBackbone=LLaMA-3.1-8B, Protocol=LongRLVR2026.0336.241.130.738.9
ShadowKVBackbone=Qwen-2.5-14B-Instruct, KV Cache Management Method=ShadowKV, KV Budget/Sparsity=20482025.0534.7940.5634.8825
RazorBackbone=Qwen-2.5-14B-Instruct, KV Cache Management Method=Razor, KV Budget/Sparsity=0.152025.0534.1943.3330.725.93
ArkValeBackbone=Qwen-2.5-14B-Instruct, KV Cache Management Method=ArkVale, KV Budget/Sparsity=20482025.0534.1941.1133.4924.07
FreeKVBackbone=Qwen-2.5-14B-Instruct, KV Cache Management Method=FreeKV, KV Budget/Sparsity=20482025.0534.1941.1133.4924.07
QuestBackbone=Qwen-2.5-14B-Instruct, KV Cache Management Method=Quest, KV Budget/Sparsity=20482025.0533.84033.4924.07
FullBackbone=Qwen-2.5-14B-Instruct, KV Cache Management Method=Full, KV Budget/Sparsity=Full Cache2025.0533.441.1131.1625
Qwen2.5-7B-1M-SFTBackbone=Qwen2.5-7B-1M, Protocol=SFT2026.0333.236.732.628.7
Qwen2.5-7B-1MBackbone=Qwen2.5-7B-1M2026.033337.831.228.7
RaaSBackbone=Qwen-2.5-14B-Instruct, KV Cache Management Method=RaaS, KV Budget/Sparsity=20482025.0532.640.5632.0920.37
LLaMA-3.1-8B-RLVRBackbone=LLaMA-3.1-8B, Protocol=RLVR2026.0332.435.631.229.6
Qwen2.5-7B-1M-RLVRBackbone=Qwen2.5-7B-1M, Protocol=RLVR2026.0332.437.229.330.6
InfiniGenBackbone=Qwen-2.5-14B-Instruct, KV Cache Management Method=InfiniGen, KV Budget/Sparsity=20482025.0532.3140.5629.8923.95
LLaMA-3.1-8B-SFTBackbone=LLaMA-3.1-8B, Protocol=SFT2026.0331.236.128.428.7
LLaMA-3.1-8BBackbone=LLaMA-3.1-8B2026.0330.434.431.621.3
FullBackbone=Llama-3.1-8B-Instruct, KV Cache Management Method=Full, KV Budget/Sparsity=Full Cache2025.0529.2234.4427.9123.15
FreeKVBackbone=Llama-3.1-8B-Instruct, KV Cache Management Method=FreeKV, KV Budget/Sparsity=20482025.0529.223527.4423.15
ArkValeBackbone=Llama-3.1-8B-Instruct, KV Cache Management Method=ArkVale, KV Budget/Sparsity=20482025.0528.6333.8926.9823.15
InfiniGenBackbone=Llama-3.1-8B-Instruct, KV Cache Management Method=InfiniGen, KV Budget/Sparsity=20482025.0528.5632.2826.0524.07
QuestBackbone=Llama-3.1-8B-Instruct, KV Cache Management Method=Quest, KV Budget/Sparsity=20482025.0528.4333.3327.4422.22
RaaSBackbone=Llama-3.1-8B-Instruct, KV Cache Management Method=RaaS, KV Budget/Sparsity=20482025.0528.2333.8926.5122.02
QuestBackbone=Qwen-2.5-7B-Instruct, KV Cache Management Method=Quest, KV Budget/Sparsity=20482025.0527.6336.6722.7922.22
RazorBackbone=Llama-3.1-8B-Instruct, KV Cache Management Method=Razor, KV Budget/Sparsity=0.152025.0527.4433.8925.1221.3
FullBackbone=Qwen-2.5-7B-Instruct, KV Cache Management Method=Full, KV Budget/Sparsity=Full Cache2025.0527.4436.1123.7220.37
ArkValeBackbone=Qwen-2.5-7B-Instruct, KV Cache Management Method=ArkVale, KV Budget/Sparsity=20482025.0526.8436.1122.3320.37
FreeKVBackbone=Qwen-2.5-7B-Instruct, KV Cache Management Method=FreeKV, KV Budget/Sparsity=20482025.0526.8434.4422.3323.15
InfiniGenBackbone=Qwen-2.5-7B-Instruct, KV Cache Management Method=InfiniGen, KV Budget/Sparsity=20482025.0526.4432.2223.2623.15
RaaSBackbone=Qwen-2.5-7B-Instruct, KV Cache Management Method=RaaS, KV Budget/Sparsity=20482025.0526.2435.5621.8619.44
LLaMA-3.1-70BBackbone=LLaMA-3.1-70B2026.0325.936.24534
ShadowKVBackbone=Qwen-2.5-7B-Instruct, KV Cache Management Method=ShadowKV, KV Budget/Sparsity=20482025.0525.8432.222026.85
ShadowKVBackbone=Llama-3.1-8B-Instruct, KV Cache Management Method=ShadowKV, KV Budget/Sparsity=20482025.0525.4532.7822.7918.52
RazorBackbone=Qwen-2.5-7B-Instruct, KV Cache Management Method=Razor, KV Budget/Sparsity=0.152025.0525.2532.7821.8619.44