Benchmark Almanac

Intern-S2-397B

Intern-S2-397B ranks first on 3 of the 12 benchmarks with five or more models scored.

Shanghai AI Lab Released September 13th, 2026 Family Intern-S2 The lab's model card Compare with others
21
Benchmarks with a score
3
Ranked first, out of five or more models
4
In the top three
Shanghai AI Lab Sep 2026

Intern-S2-397B

Open weights Flagship

154.1 #66 of 404
Self-Reported Index, from the lab's own numbers. No Epoch Capabilities Index score yet.
  • MMMU-Pro 81.7% #9 /90
  • SciCode 47.6% #14 /49
  • ChartQA Pro 71.1% #3 /8

Where it ranks

Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.

Knowledge and language

BenchmarkScoreRankSource
MMLU-Pro 89.8% #1 of 167 Lab model card
AdvancedIF 71.6% #5 of 6 Lab model card

Math and science

BenchmarkScoreRankSource
FrontierScience 87% #1 of 8 Lab model card
HMMT 2026 93.6% #6 of 19 Lab model card
SciReasoner evaluation suite 63.3% #1 of 2 Lab model card
IMO-ProofBench 38.4% #5 of 5 Lab model card
AdvancedMathBench 58.1% #1 of 1 Lab model card
SGI-Bench 51.9% #1 of 1 Lab model card

Coding

BenchmarkScoreRankSource
SWE-bench Multilingual 84% #9 of 73 Lab model card
SWE-Bench Pro 68.5% #10 of 81 Lab model card
SciCode 47.6% #14 of 49 Lab model card

Agents and tools

BenchmarkScoreRankSource
Terminal-Bench 2.1 64% #42 of 56 Lab model card

Vision, audio and robotics

BenchmarkScoreRankSource
MMMU-Pro 81.7% #9 of 90 Lab model card
Scientists' First Exam (SFE) 60.7% #1 of 5 Lab model card
MicroVQA 69.7% #1 of 3 Lab model card
ChartQA Pro 71.1% #3 of 8 Lab model card
XLRS-Bench 51.4% #4 of 4 Lab model card

Safety and security

BenchmarkScoreRankSource
Biology-Instructions 55.7% #1 of 2 Lab model card
Mol-Instructions 54% #1 of 2 Lab model card
MolecularIQ 62.4% #1 of 1 Lab model card
TOMG-Bench 66.8% #1 of 1 Lab model card

Its lab's numbers put it at 154.1, number 66 of 404

The Self-Reported Index fits one score to the 9 capability results Shanghai AI Lab reported for this model, on the scale of the Epoch Capabilities Index. Labs choose which benchmarks to show and how to run them, so read it as the lab's claim, not an independent measurement. The 90% interval runs from 149.9 to 158.2. How we compute it.

Its launch post

The benchmarks the lab chose to headline, with the scores it gave.

Intern-S2-397B model card (Hugging Face) Sep 13, 2026