Benchmark Almanac

Falcon H1R 7B

Falcon H1R 7B.

Technology Innovation Institute Released January 5th, 2026 Family Falcon-H1 The lab's model card Compare with others
18
Benchmarks with a score
0
Ranked first, out of five or more models
0
In the top three
Technology Innovation Institute Jan 2026

Falcon H1R 7B

Open weights Small

133.5 #244 of 404
Self-Reported Index, from the lab's own numbers. No Epoch Capabilities Index score yet.
  • IFBench 53.4% #62 /80
  • SciCode 28.3% #45 /49
  • Terminal-Bench Hard 4.96% #12 /13

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 72.1% #117 of 167 Lab model card
IFBench 53.4% #62 of 80 Lab model card
Humanity's Last Exam 11.1% #135 of 154 Lab model card

Math and science

BenchmarkScoreRankSource
MATH-500 97.4% #16 of 70 Lab model card
AIME 2024 88.1% #23 of 88 Lab model card
SMT 2025 85.8% #18 of 40 Official leaderboard
AIME 2025 83.1% #80 of 143 Lab model card
GPQA Diamond 61.3% #217 of 302 Lab model card
BRUMO 2025 85.8% #31 of 42 Official leaderboard
CMIMC 2025 71.9% #26 of 33 Official leaderboard
HMMT 2025 64.9% #50 of 54 Lab model card
HMMT Nov 2025 79.2% #21 of 22 Official leaderboard
AMO-Bench 36.3% #4 of 4 Lab model card

Coding

BenchmarkScoreRankSource
LiveCodeBench v6 68.6% #60 of 90 Lab model card
SciCode 28.3% #45 of 49 Lab model card

Agents and tools

BenchmarkScoreRankSource
Terminal-Bench Hard 4.96% #12 of 13 Lab model card

Safety and security

BenchmarkScoreRankSource
WildJailbreak 98.5% #1 of 2 Lab model card
XSTest 96.4% #15 of 16 Lab model card

Its lab's numbers put it at 133.5, number 244 of 404

The Self-Reported Index fits one score to the 11 capability results Technology Innovation Institute 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 131.2 to 135.7. How we compute it.

Its lab's numbers next to 2 leaderboards

Where the benchmark's own leaderboard, or Epoch AI's run, also has this model. Gaps mostly come from different harnesses, effort levels and task subsets. Every pair.

Benchmark Lab says Board says Gap
HMMT 2025 64.9% HMMT-25 84.2% Leaderboard −19 pts
AIME 2025 83.1% AIME-25 86.7% Leaderboard −3.6 pts

Its launch post

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

Its system card

Every result in the longer documents the lab published with it. 7 of them aren't in the launch post.