Coding on WildChat 5,000 conversations benchmark leaderboard · SOTA2 Research

Coding on WildChat 5,000 conversations

0.26KL Divergence (Forward)Gemini 3.1 Pro

Evaluation Results

Method	Links
Gemini 3.1 Pro2026.05		0.26	0.256	0.246
GPT-5.42026.05		0.261	0.265	0.248
humanlm-opinion2026.05		0.268	0.257	0.247
gemma-4-31B-it2026.05		0.311	0.296	0.297
UserLM-8b2026.05		0.328	0.388	0.349
GPT-5.4 nano2026.05		0.353	0.364	0.353
Gemini 3 Flash2026.05		0.355	0.328	0.334
gpt-oss-120b2026.05		0.358	0.353	0.343
GPT-5.4 mini2026.05		0.363	0.388	0.364
Qwen3.5-122B-A10B2026.05		0.394	0.378	0.387
gemma-4-26B-A4B-it2026.05		0.397	0.373	0.379
gpt-oss-20b2026.05		0.411	0.435	0.409
gemma-4-E4B-it2026.05		0.422	0.383	0.402
Gemini 3.1 Flash-Lite2026.05		0.432	0.406	0.421
Llama-3.1-8B-Instruct2026.05		0.435	0.427	0.423
Qwen3.5-27B2026.05		0.437	0.421	0.433
Claude Haiku 4.52026.05		0.438	0.43	0.427
gemma-4-E2B-it2026.05		0.446	0.435	0.435
Qwen3.5-9B2026.05		0.465	0.457	0.454
Llama-3.3-70B-Instruct2026.05		0.471	0.456	0.469
Qwen3.5-35B-A3B2026.05		0.48	0.467	0.478
Qwen3.5-4B2026.05		0.482	0.471	0.472
Qwen3.5 .8B2026.05		0.547	0.568	0.579
Qwen3.5-2B2026.05		0.584	0.591	0.608