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

Discourse-level Chinese-English translation on DiscoX

42.62Accuracy

GPT-5

Updated 13d ago

Evaluation Results

Method	Links
GPT-5 2026.04		42.62	12.42	12.4
Gemini2.5-Pro 2026.04		40.23	11.86	13.26
Claude-4-sonnet 2026.04		38.65	11.42	11.8
Qwen3-8B + R2-Write-SFT + RLp 2026.04		23.5	10.82	10.9
Qwen3-8B + R2-Write-SFT + RL 2026.04		23.16	10.91	10.95
Qwen3-8B + R2-Write-SFT 2026.04		22.7	10.06	10.9
Qwen3-8B + Distill 2026.04		21.7	9.3	10
Qwen3-8B + Distill + RL 2026.04		20.42	9.84	10.2
Qwen3-8B + R2-Write-Last 2026.04		20.14	9.2	9.24
Qwen3-8B + RL 2026.04		19.3	9.53	9.95
LongWriter-Zero-32B 2026.04		19.2	9.74	9.2
Qwen3-8B 2026.04		18.8	8.74	8.8
Qwen3-4B + R2-Write-SFT + RLp 2026.04		18.64	10.22	10.28
Reverse-Engineering-8B 2026.04		18.62	9.58	9.86
Qwen3-4B + R2-Write-SFT 2026.04		18.22	8.48	9.4
Qwen3-4B + RL 2026.04		15.26	7.86	8.96
Writing-RL-7B 2026.04		15.2	7.7	7.2
Longwriter-9B 2026.04		14.12	7.2	7.88
Qwen3-4B 2026.04		12.82	7.7	8.2