QuantID/Measurement Lab

Measuring AI, science and risk with reproducible methods

Hugging Face Official Benchmarks: Daily Update 2026-11-19

2026-11-19 · QuantID Measurement Lab

aibenchmarkshuggingfacellm

This is the daily update of the Hugging Face Official Benchmarks map, which reads every Hugging Face official benchmark leaderboard straight from the public API.

As of 2026-11-19 (KST): 63 official benchmarks, 1,462 leaderboard entries.

What changed since the last snapshot

  • safeagents-judge-bench: new #1 PromptforceAI/niti-gate-vegaml-4b (score 0.905), previously vllm-sr/d3-ultra.
  • SWE-bench_Pro: new #1 thinkingmachines/Inkling-2-Small (score 97.7), previously PCL-RSI-lab/Seshat-agent.
  • terminal-bench: new #1 zai-org/GLM-5.3 (score 41.2), previously moonshotai/Kimi-K3.1-Pro.
  • SWE-bench_Multimodal: new #1 thinkingmachines/Inkling-2-Small (score 43.1), previously no entries.
  • New official benchmarks: ParseBench-2, multiloko, IFBench_test.
  • 33 new leaderboard entries across all boards.

Organizations by number of #1 spots

Rank Organization #1 spots
1 FINAL-Bench 11
2 moonshotai 7
3 zai-org 6
4 Qwen 3
5 deepseek-ai 3
6 KDLAI 2
7 MiniMaxAI 2
8 XiaomiMiMo 2
9 inclusionAI 2
10 ornith-ai 2

Current #1 on every official benchmark

Benchmark #1 model Score Entries
MMLU-Pro internlm/Intern-S2-Preview-397B 89.75 163
gpqa FINAL-Bench/Darwin-180B-RSI 94.44 124
hle moonshotai/Kimi-K3.1-Pro 70 104
SWE-bench_Verified ornith-ai/Ornith-1.5-397B 86 78
longshot-bench Qwen/Qwen3-Omni-30B-A3B-Thinking 64.2 67
open-asr-leaderboard Edge0/ARK-ASR-3B 4.76 65
SWE-bench_Pro thinkingmachines/Inkling-2-Small 97.7 56
WildClawBench Qwen/Qwen3.8-2.4T-A95B 708 52
terminal-bench-2.1 deepseek-ai/DeepSeek-V4.1-Flash 90.6 45
terminal-bench-2.0 mindlab-research/Macaron-V1-Venti 87.6 44
aime_2026 deepseek-ai/DeepSeek-V4.1-Flash 100 39
ParseBench KDLAI/KDL-Frontier-Parser-nano-v2 78.95935214038234 35
SWE-bench_Multilingual inclusionAI/Ling-3.1-flash 88.5 35
deep-swe PCL-RSI-lab/Seshat-agent 90.3 34
ScreenSpot-Pro beipo-ai/NorthCUA-Pro 93.3 33
gsm8k XiaomiMiMo/MiMo-V2.5-Pro 99.6 28
hmmt_feb_2026 FINAL-Bench/Darwin-180B-RSI 100 26
OmniCompass moonshotai/Kimi-K3.1-Pro 43.5 26
olmOCR-bench SoMarkAI/OXR-1.0 87.84 24
IFBench_test Qwen/Qwen3.8-2.4T-A95B 82.8 23
Claw-Eval ornith-ai/Ornith-1.5-397B 81.4 22
ParseBench-2 KDLAI/KDL-Frontier-Parser-nano-v2 78.89 20
MDPBench FINAL-Bench/Darwin-180B-RSI 83.65 19
safeagents-judge-bench PromptforceAI/niti-gate-vegaml-4b 0.905 18
Toolathlon XiaomiMiMo/MiMo-V2.6-Pro-RL 76.9 18
Real5-OmniDocBench PaddlePaddle/PaddleOCR-VL-1.6 93.19 18
RedlineBench crosbylegal/gemini-4-argon 62.7 17
VDEBench sensenova/SenseNova-U1.5-8B-MoT 0.828 16
MMMU_Pro moonshotai/Kimi-K3.1-Pro 90 14
apex-agents moonshotai/Kimi-K3.1-Pro 69.4 14
ExtractBench FINAL-Bench/Darwin-180B-RSI-R3 90.29 12
skillsbench inclusionAI/Ling-3.1-flash 68.7 12
LEXam-hard FINAL-Bench/Darwin-180B-RSI 45.72 12
Vaani-Benchmark-V1.0 gnani/gnani 13 11
agents-last-exam moonshotai/Kimi-K3.1-Pro 43.2 10
terminal-bench-science moonshotai/Kimi-K3.1-Pro 38.1 10
mmi meta-models/Muse-Spark-Open-816B-A42B 15.55 10
LEXam FINAL-Bench/Darwin-180B-RSI 68.94 10
yc-bench MiniMaxAI/MiniMax-M3 2100000 10
physical-engineerings-last-exam zai-org/GLM-5.2 26 9
typed-decisions FINAL-Bench/Darwin-27B-ZTC-v5 0.812 9
Long-Horizon-Terminal-Bench moonshotai/Kimi-K3 6 9
multiloko meta-llama/Llama-3.1-405B 34.31 8
WBench robbyant/lingbot-world-fast 78.8 8
ResearchClawBench zai-org/GLM-5.2 20.709230769230768 7
Video-MME-v2 MiniMaxAI/MiniMax-M3 85.4 7
ifstruct-v1.0 FINAL-Bench/Darwin-180B-RSI 98.95 6
EvasionBench zai-org/GLM-4.7 82.91 5
cwe-bench deepseek-ai/DeepSeek-V4.1-Flash-1017 57 4
arguana FINAL-Bench/Darwin-180B-RSI-R3 83.83 4
chi-bench zai-org/GLM-5.1 18.7 3
BRIGHT FINAL-Bench/Darwin-BRIGHT-Retriever 0.307 2
terminal-bench zai-org/GLM-5.3 41.2 2
APEX-v1-extended zai-org/GLM-4.7 51.7 2
decision-index-leaderboard FINAL-Bench/Darwin-27B-ZTC 63.03 1
SWE-bench_Multimodal thinkingmachines/Inkling-2-Small 43.1 1
vlabench_primitive_ft_lerobot_video nvidia/GR00T-N1-2B 39.7 1

How to read this

Most leaderboard entries are self-reported by model authors, and evaluation settings differ between submissions (sampling, voting, thinking budget). An entry is marked verified only where Hugging Face marks it so. Ties at rank 1 count for each tied organization.

Disclosure: QuantID is a technology partner of VIDRAFT, whose FINAL-Bench organization appears in these rankings. This post reports the public leaderboard data as published on Hugging Face; it is not an independent evaluation.