Reasoning with Reka Flash 3 (Reka news) 10 Mar 2025
- GPQA 61.1
- HumanEval 96.3
- LiveCodeBench v5 62
- MATH-500 90.6
- AIME'24 (cons@64) 70
- AIME'25 (cons@64) 50
- MMLU-Pro 65
- BFCL v2 77.2
Reka Flash 3
Reka Flash 3 places in the top three on HumanEval.
Reka Released 10 Mar 2025
13
Benchmarks with a score
0
Ranked first, out of five or more models
1
In the top three
Reka Mar 2025
Reka Flash 3
Open weights
No Epoch Capabilities Index score yet
- CritPt 0% #124 /160
- LiveCodeBench v5 62% #23 /44
- BALROG 29.2% #18 /35
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
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Creative Writing v3 | 54.8 | #97 of 120 | Official leaderboard | |
| MMLU-Pro | 65% | #123 of 151 | Model card | |
| Longform Creative Writing (EQ-Bench) | 41.2 | #98 of 119 | Official leaderboard |
Math and science
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| GPQA Diamond | 61.1% | #206 of 284 | Model card | |
| AIME 2024 | 70% | #64 of 87 | Model card | |
| MATH-500 | 90.6% | #51 of 67 | Model card | |
| CritPt | 0% | #124 of 160 | Epoch AI | |
| AIME 2025 | 50% | #131 of 143 | Model card |
Coding
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| HumanEval | 96.3% | #3 of 95 | Model card | |
| LiveCodeBench v5 | 62% | #23 of 44 | Model card |
Agents and tools
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Berkeley Function Calling Leaderboard (BFCL) | 77.2% | #6 of 47 | Model card | |
| BALROG | 29.2% | #18 of 35 | Official leaderboard | |
| SnakeBench | 11 | #176 of 183 | Official leaderboard |
Its launch post
The benchmarks the lab chose to headline, with the scores it gave.