DForkTree Study - 128-Prompt Greedy AR - GSM8K: leaderboard

DForkTree v1 Table7,128promptsubset ofGSM8K,notfullbenchmark. Unchanged autoregressivenonthinkingBF16,Qwen3.5,SGLang0.5.22,RTX5090,batch1,temp0,max2048tokens. GSMlastnumber;MATHlastcompleteboxednormalized;HumanEvalofficialtests;MBPPfixedfunctionnamealiaswhenpromptomitsname. No copiedaccuracyunderdecoderchanges or throughputmetrics.

Metric: Score (%). Source: arxiv.org. Saturation forecast: Around February 2027. 2 models tracked.

Top models

#ModelScoreOverall rank
1Qwen 3.5 9B94.5#803
2Qwen 3.5 4B91.4#1180

No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.

Interactive version: theaggregate.ai/benchmark?slug=dforktree-study-128-prompt-greedy-ar-gsm8k · How It Works · Data refreshed daily, snapshot 2026-11-12.