Ember-1
Ember-1.
14
Benchmarks with a score
0
Ranked first, out of five or more models
0
In the top three
Fireworks AI Sep 2026 Ember-1
Closed weights Flagship
No Epoch Capabilities Index score yet
- Harvey LAB (Legal Agent Benchmark) 9.17% #23 /80
- Vals Finance Agent v2 51.9% #38 /80
- Legal Research Bench 44.7% #22 /88
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.
Math and science
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| ProofBench v1.1 | 61% | #22 of 51 | Official leaderboard | |
| MysteryMechanism | 22.5% | #20 of 29 | Official leaderboard |
Coding
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Vibe Code Bench | 83.9% | #21 of 115 | Official leaderboard | |
| Code Migration | 32.5% | #39 of 78 | Official leaderboard | |
| IOI v2 (Vals) | 46.3% | #31 of 45 | Official leaderboard |
Professional work
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Legal Research Bench | 44.7% | #22 of 88 | Official leaderboard | |
| Harvey LAB (Legal Agent Benchmark) | 9.17% | #23 of 80 | Official leaderboard | |
| Tax Agent Bench | 31.1% | #26 of 79 | Official leaderboard | |
| Public Benefits Bench | 66.2% | #19 of 50 | Official leaderboard | |
| Excel Modeling Benchmark | 63.4% | #29 of 76 | Official leaderboard | |
| Vals Finance Agent v2 | 51.9% | #38 of 80 | Official leaderboard |
Safety and security
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| CyberBench v1.1 (Vals) | 69.5% | #21 of 43 | Official leaderboard |
Other
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| SAGE v1.1 (Vals) | 86.4% | #12 of 28 | Official leaderboard | |
| Vals Science Index | 31.8% | #16 of 23 | Official leaderboard |