Benchmark Almanac

Flint

Flint.

15
Benchmarks with a score
0
Ranked first, out of five or more models
0
In the top three
Recursive Superintelligence

Flint

Open weights Small

147.0 #122 of 412
Self-Reported Index, from the lab's own numbers. No Epoch Capabilities Index score yet.
  • The Aggregate Day-one Recall Check 6.20% #33 /36
  • ReactBench 17.9% #27 /29
  • The Aggregate Puzzle Portfolio Ladder 66.7% #19 /24

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
The Aggregate Day-one Recall Check 6.20% #33 of 36 Official leaderboard
IFBench 89.2% Not ranked Lab model card
IFEval 95.1% Not ranked Lab model card

Coding

BenchmarkScoreRankSource
DesignArena UI Component 1,157 #88 of 149 Official leaderboard
DesignArena Website 1,151 #103 of 162 Official leaderboard
DesignArena Code 1,137 #103 of 153 Official leaderboard
DesignArena Game Dev 1,075 #117 of 154 Official leaderboard
ReactBench 17.9% #27 of 29 Official leaderboard
LiveCodeBench v6 80.3% Not ranked Lab model card
SWE-Bench Pro V2 89.7% Not ranked Lab model card
SWE-bench Verified 74.2% Not ranked Lab model card

Agents and tools

BenchmarkScoreRankSource
The Aggregate Puzzle Portfolio Ladder 66.7% #19 of 24 Official leaderboard
The Aggregate Inversion and Execution Ladder 54.5% #16 of 20 Official leaderboard
Terminal-Bench 2.1 54.9% Not ranked Lab model card

Professional work

BenchmarkScoreRankSource
BRIDGE 47.1 #22 of 48 Official leaderboard

Its lab's numbers put it at 147.0, number 122 of 412

The Self-Reported Index fits one score to the 5 capability results Recursive Superintelligence 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 142.7 to 151.3. How we compute it.

Who measured it?

Independent runs cover 9 of its 15 benchmarks. For the other 6, we have only Recursive Superintelligence's own number or a claim Epoch AI collected.

  • An independent run only 9
  • A claim only 6
Official leaderboards scored it on 9.

Sources: Official leaderboards, Epoch AI (CC BY 4.0) and labs’ model cards

Its launch post

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