Reka AIproprietary
Released: 2024-04-15Reka Core
Reka multimodal frontier model processing text, images, video, and audio with native multilingual reasoning.
Intelligence Index
79.5/ 100
Calibrated multi-domain compositeThroughput Speed
55tok/sec
Output streaming throughputTime To First Token
440ms
Initial chunk server latencyLiveBench Contam-Free
69.2%
September 2026 suite scoreSourced Benchmark Evaluations (23)
Verified performance across authoritative benchmarks with provenance tracking.| Benchmark | Status | Score | Trust | Date | Source Type | Provenance |
|---|---|---|---|---|---|---|
| LiveBench (Continuous) | Active | 69.2% | 94 | 2025-02-01 | independent | Source ↗ |
| StructFlowBench | Nearing Saturation | 62.2% | 75 | 2024-05-03 | independent | Source ↗ |
| FireBench | Active | 61.4% | 95 | 2024-05-01 | independent | Source ↗ |
| SimpleQA | Active | 41.3% | 90 | 2024-04-30 | independent | Source ↗ |
| OpenAI-MRCR | Active | 66.8% | 80 | 2024-04-30 | independent | Source ↗ |
| ComplexBench | Nearing Saturation | 60% | 80 | 2024-04-29 | independent | Source ↗ |
| LongBench-Pro | Active | 65.2% | 80 | 2024-04-29 | independent | Source ↗ |
| ZeroSCROLLS | Nearing Saturation | 63.6% | 78 | 2024-04-28 | independent | Source ↗ |
| InfiniteBench | Nearing Saturation | 67.6% | 65 | 2024-04-27 | independent | Source ↗ |
| AgentIF | Nearing Saturation | 49.7% | 80 | 2024-04-27 | independent | Source ↗ |
| MMLU-Pro | Nearing Saturation | 73.1% | 75 | 2024-04-26 | independent | Source ↗ |
| SCROLLS | Nearing Saturation | 63.6% | 65 | 2024-04-26 | independent | Source ↗ |
| RULER | Nearing Saturation | 73.1% | 84 | 2024-04-25 | independent | Source ↗ |
| MMLU-Redux | Saturated | 77.9% | 25 | 2024-04-24 | independent | Source ↗ |
| QuALITY | Nearing Saturation | 70% | 78 | 2024-04-24 | independent | Source ↗ |
| LongBench | Nearing Saturation | 68.4% | 65 | 2024-04-23 | independent | Source ↗ |
| QASPER | Nearing Saturation | 66% | 65 | 2024-04-22 | independent | Source ↗ |
| TruthfulQA | Saturated | 70% | 25 | 2024-04-22 | independent | Source ↗ |
| MT-Bench | Nearing Saturation | 7.5% | 75 | 2024-04-20 | independent | Source ↗ |
| IFEval | Deprecated | 71.7% | 25 | 2024-04-19 | independent | Source ↗ |
| GPQA | Nearing Saturation | 70% | 78 | 2024-04-19 | independent | Source ↗ |
| Needle-in-a-Haystack | Saturated | 95.9% | 44 | 2024-04-18 | independent | Source ↗ |
| MMLU | Saturated | 81.1% | 25 | 2024-04-16 | independent | Source ↗ |
OpenAI Compatible API ExecutionAPI Ready
from openai import OpenAI
client = OpenAI()
response = client.chat.completions.create(
model="reka-core",
messages=[{"role": "user", "content": "Evaluate multi-step logic problem"}],
temperature=0.2
)
print(response.choices[0].message.content)Evaluation & Sourcing Notes
Scores listed for Reka Core represent verified evaluations extracted from official research papers, independent evaluation suites (HELM, LMSYS, OpenCompass, LiveBench), and verified audit reports.
All benchmarks marked as saturated or deprecated reflect historical performance where the benchmark no longer provides active discriminative power.