Long-context Reasoning on AA-LCR (Accuracy and Avg.)
81AccuracyQwen3.5-9B
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Qwen3.5-9BQuantization=BF162026.05 | 81 | 72.04 | — | |
| Mix-QuantBase Model=Qwen3.5-9B2026.05 | 79.33 | 70.59 | — | |
| Qwen3.5-9B-NVFP4Quantization=NVFP42026.05 | 78 | 63.26 | — | |
| Gemma-4-31B-itQuantization=BF162026.05 | 76.67 | 82.36 | — | |
| Mix-QuantBase Model=Gemma-4-31B-it2026.05 | 73.67 | 81.39 | — | |
| Gemma-4-31B-it-NVFP4Quantization=NVFP42026.05 | 71.67 | 78.26 | — | |
| Gemma-4-26B-A4B-itQuantization=BF162026.05 | 67 | 71.94 | — | |
| Mix-QuantBase Model=Gemma-4-26B-A4B-it2026.05 | 64.33 | 71.93 | — | |
| Gemma-4-26B-A4B-it-NVFP4Quantization=NVFP42026.05 | 50.67 | 66.31 | — | |
| Qwen3-8BQuantization=BF162026.05 | 33.67 | 62.11 | — | |
| Mix-QuantBase Model=Qwen3-8B2026.05 | 28.67 | 61 | — | |
| Qwen3-8B-NVFP4Quantization=NVFP42026.05 | 24.67 | 55.22 | — | |
| Qwen3-30B-A3B-Thinking-2507Base Model=Qwen3-30B-A3B-Thinking-2507, Training Recipe=Base2026.06 | — | — | 48.75 | |
| Qwen3-30B-A3B-Thinking-2507 + Data Recipe for Long-Context RLBase Model=Qwen3-30B-A3B-Thinking-2507, Training Recipe=Data Recipe for Long-Context RL2026.06 | — | — | 56 | |
| Qwen3-30B-A3B-Thinking-2507 + DocQA-RL-1.6KBase Model=Qwen3-30B-A3B-Thinking-2507, Training Recipe=DocQA-RL-1.6K2026.06 | — | — | 43.3 | |
| Qwen3-30B-A3B-Thinking-2507 + KeyChain-15KBase Model=Qwen3-30B-A3B-Thinking-2507, Training Recipe=KeyChain-15K2026.06 | — | — | 45.25 | |
| Qwen3-4B-Thinking-2507Base Model=Qwen3-4B-Thinking-2507, Training Recipe=Base2026.06 | — | — | 35.25 | |
| Qwen3-4B-Thinking-2507 + Data Recipe for Long-Context RLBase Model=Qwen3-4B-Thinking-2507, Training Recipe=Data Recipe for Long-Context RL2026.06 | — | — | 45.75 | |
| Qwen3-4B-Thinking-2507 + DocQA-RL-1.6KBase Model=Qwen3-4B-Thinking-2507, Training Recipe=DocQA-RL-1.6K2026.06 | — | — | 32.5 | |
| Qwen3-4B-Thinking-2507 + KeyChain-15KBase Model=Qwen3-4B-Thinking-2507, Training Recipe=KeyChain-15K2026.06 | — | — | 41 | |
| Qwen3-8B-128KBase Model=Qwen3-8B-128K, Training Recipe=Base2026.06 | — | — | 28.75 | |
| Qwen3-8B-128K + Data Recipe for Long-Context RLBase Model=Qwen3-8B-128K, Training Recipe=Data Recipe for Long-Context RL2026.06 | — | — | 34.5 | |
| Qwen3-8B-128K + DocQA-RL-1.6KBase Model=Qwen3-8B-128K, Training Recipe=DocQA-RL-1.6K2026.06 | — | — | 28.5 | |
| Qwen3-8B-128K + KeyChain-15KBase Model=Qwen3-8B-128K, Training Recipe=KeyChain-15K2026.06 | — | — | 34.5 |