List your company

Long-context Reasoning on FRAMES

84.7ScoreClaude-Opus-4.5
69.72473.61277.581.388May 29, 2026Jun 3, 2026Jun 8, 2026Jun 14, 2026Jun 19, 2026Jun 24, 2026Jun 30, 2026

Evaluation results

MethodLinks
Claude-Opus-4.52026.0684.7
Seed2.0 Pro2026.0684.5
GPT-5.2 High2026.0684
Gemini-3-FlashTier=High2026.0683.7
DocQABackbone=Qwen3-30B-A3B-Thinking-25072026.0583.5
Seed2.0 Lite2026.0683.4
GPT-5-miniTier=High2026.0682.9
LoongRLBackbone=Qwen3-30B-A3B-Thinking-25072026.0581.9
LONGTRACERLBackbone=Qwen3-30B-A3B-Thinking-25072026.0581.9
Gemini-3-pro High2026.0681.9
LongRLVRBackbone=Qwen3-30B-A3B-Thinking-25072026.0581.7
BaseBackbone=Qwen3-30B-A3B-Thinking-25072026.0580.7
Seed2.0 Mini2026.0680.5
LONGTRACERL-GRPOBackbone=Qwen3-30B-A3B-Thinking-25072026.0579.6
LONGTRACERLBackbone=Qwen3-4B-Thinking-25072026.0579.5
Claude-Sonnet-4.52026.0678.7
LongRLVRBackbone=Qwen3-4B-Thinking-25072026.0578.5
DocQABackbone=Qwen3-4B-Thinking-25072026.0578.3
BaseBackbone=Qwen3-4B-Thinking-25072026.0576.7
LONGTRACERL-GRPOBackbone=Qwen3-4B-Thinking-25072026.0576.1
LoongRLBackbone=Qwen3-4B-Thinking-25072026.0575.8
LONGTRACERLBackbone=DeepSeek-R1-0528-Qwen3-8B2026.0574.3
DocQABackbone=DeepSeek-R1-0528-Qwen3-8B2026.0573.4
BaseBackbone=DeepSeek-R1-0528-Qwen3-8B2026.0573.2
LONGTRACERL-GRPOBackbone=DeepSeek-R1-0528-Qwen3-8B2026.0573.1
LoongRLBackbone=DeepSeek-R1-0528-Qwen3-8B2026.0572.6
LongRLVRBackbone=DeepSeek-R1-0528-Qwen3-8B2026.0570.3