JFinBench - Tools - Terminus-2 Harbor 0.24.0: leaderboard

Same321examplesandGPT5.1grading, mean3runs; terminal/webtools; fixednamedharness; highestnativeeffort. Allcompleted/failedattemptsperAppendixD, notbest-ofharnessselection. PreserveDS1017checkpoint; noSPRO/reasoningdefaults inferred. Metric: Overall score (%). Source: https://arxiv.org/html/2611.18130v1

Metric: Overall score (%). Source: arxiv.org. Saturation forecast: Estimated already saturated. 12 models tracked. Updated 17 Nov 2026.

Top models

#ModelScoreOverall rank
1GPT-6 Astra97.3#8
2Claude Fable 596.4#28
3Gemini 3.8 Flash94.2#39
4Kimi K3.1 Pro92.1#32
5DeepSeek V4.1 Flash (1017)88.6#44
6GLM-5.3 Flash87.6#136
7Kimi K386.1#59
8MiMo-V2.6-Pro85.3#86
9Qwen 3.8 Flash83.3#128
10MiMo-V2.6-Flash73.2#176
11Kimi K2.670.4#191
12Mistral Large 470.1#143

No result here: #1 Claude Fable 5.2, #2 GPT-6.2 Astra, #3 Gemini 4 Carbon.

Interactive version: theaggregate.ai/benchmark?slug=jfinbench-tools-terminus-2-harbor-0-24-0 · How It Works · Data refreshed daily, snapshot 2026-11-17.