LongBench v2 — Results

Rank	Model	Score
1	ling-3-0-flash-base	52.62
2	deepseek-v4-1-flash	45.2
3	ling-3-0-tiny-base	43.85
4	minicpm5-2b	43.7
5	qwen3-5-397b-a17b	0.632
6	qwen3-6-plus	0.62
7	nemotron-3-ultra-550b-a55b	0.619
8	minimax-m1-80k	0.615
9	kimi-k2-5	0.61
10	mai-thinking-1	0.61
11	minimax-m1-40k	0.61
12	mimo-v2-flash	0.606
13	qwen3-5-27b	0.606
14	qwen3-5-122b-a10b	0.602
15	qwen3-5-35b-a3b	0.59
16	qwen3-5-9b	0.552
17	qwen3-5-4b	0.5
18	deepseek-v3	0.487
19	qwen3-5-2b	0.387
20	qwen3-5-08b	0.261

LongBench v2 — Evaluations

1 phaseActive

No evaluations yet for this benchmark.