IndicBankBench - Long Context: leaderboard

Metric: Strict pass^3 (%; 89 cases passed out of the 89 cases of the long-context axis; each case run three times at temperature 0.7 with the stated reasoning modes; a case passes only when all three runs pass the deterministic safety and tool-use gates and the GLM-5.2 response judge; scripted Indian retail-banking conversations over deterministic mock tools). Source: arxiv.org. Saturation forecast: Around 2031. 11 models tracked.

Top models

#ModelScore
1Gemma 4 31B (IT) (Thinking)60.67
2Grok 4.6 (High)55.06
3Gemini 3.8 Flash52.81
4Claude Opus 5 (Max)52.81
5DeepSeek V4 Pro (0813) (Max)52.81
6GLM-5.3 Flash (High)48.31
7DeepSeek V4 Flash (0731) (Max)44.94
8gemma-4-26B-A4B-it (Thinking)44.94
9Gemma 4 E4B (IT) (Thinking)44.94
10MiniMax-M334.83

Interactive version: theaggregate.ai/benchmark?slug=indicbankbench-long-context · How It Works · Data refreshed daily, snapshot 2026-09-26.