Kolibri-1 vs Ripple-65B Preview
New ComparisonScores are fetched live from the current database on every visit — nothing is cached.
Winner
Ripple-65B Preview
Ripple-65B Preview wins this comparison with an overall score of 76 vs 68 for Kolibri-1, leading 5 of 7 shared benchmarks.
Ripple-65B Preview
76.4
Kolibri-1
67.5
Based on 7 of 63 benchmarks scored by all 2 selected models.
Benchmark Scores Comparison
Kolibri-1
Ripple-65B Preview
Kolibri-1
78.1B params
Ripple-65B Preview
65.0B params
Coverage matrix
2 models × 63 benchmarks — hover a benchmark name for its category
Shading = score relative to the best score among the selected models:
Low
Mid
High
Missing (—)
The totals row shows how many of the 63 benchmarks each model has scores for.
| Benchmark | Kolibri-1 Aleph Alpha | Ripple-65B Preview Ricursive Intelligence |
|---|---|---|
| OmniScience Accuracy | 14.8 | — |
| GPQA Diamond | 84.3 | 85.2 |
| IFBench (prompt loose) | 78.1 | — |
| SealQA (12 distractors, <24k) | 61 | — |
| LiveCodeBench v6 | 85.9 | — |
| SWE-bench Pro | — | 61.42 |
| BFCL-V4 | 61.4 | — |
| AA-LCR | 68.3 | 75 |
| AIME 26 | 96 | 95 |
| FRAMES (<24k) | 71.2 | — |
| German Public Sector | 75 | — |
| OmniScience Non-Hallucination | 44 | — |
| Aerospace | 58.9 | — |
| Agentic Wiki QA (DE) | 69.4 | — |
| AIME 2025 (DE) | 87.5 | — |
| AutomationBench (Public) | — | 36.4 |
| Overall (EN) | 75.5 | — |
| StrongREJECT v2 (adaptive) | — | 97.5 |
| MuSiQue (EN) | 77.3 | — |
| SWE-bench Pro v2.0 Hard | — | 79.1 |
| SealQA (no distractors, >24k) | 100 | — |
| HumanEval+ | 92.7 | — |
| MMLU-ProX | 75.5 | — |
| MMLU-Pro | — | 84.2 |
| AIME 2024 | — | 97.1 |
| BrowseComp | 29.4 | — |
| BFCL v4 (web search) | 62.5 | — |
| RGB Closed-Book | 51 | — |
| Industrial Drive Technology | 60 | — |
| AA-Omniscience Index (public set) | -32.8 | — |
| MMLU | — | 89.6 |
| Humanity's Last Exam (DE) | 15.9 | — |
| IFBench | — | 79.2 |
| SWE-bench Verified | 66.4 | 76.5 |
| Humanity's Last Exam | 21.5 | — |
| BFCL v4 (non-live AST) | 79.1 | — |
| BFCL v4 (live) | 78.9 | — |
| BFCL v4 (memory) | 62.8 | — |
| AIME 2026 (DE) | 90 | — |
| Terminal-Bench 2.1 (Terminus-2) | — | 80.5 |
| Automotive Supplier | 99 | — |
| GPQA Diamond (DE) | 81.3 | — |
| HMMT Feb 26 | — | 90.5 |
| SQuAD (Utility Accuracy) | 83.4 | — |
| Terminal Bench 2.1 | 27.7 | 80.5 |
| AIME 2025 | 96.9 | 98.3 |
| BFCL v4 (multi-turn) | 47.5 | — |
| RGB Fact-Check (Error Correction) | 34 | — |
| SQuAD (M/A Grounding Score) | 23.4 | — |
| LongBenchPro | 64.5 | — |
| FRAMES (>24k) | 73 | — |
| Semiconductors | 80.4 | — |
| Honeypot | 80.8 | — |
| TauBench V3 Banking | 38.1 | 31.27 |
| MCP-Atlas | — | 71.67 |
| DeepSWE 1.1 | — | 44.3 |
| FORTRESS Benign | — | 95.5 |
| RGB Negative (Abstention) | 85.6 | — |
| MMLU Pro COT | 80 | — |
| Overall (DE) | 70.8 | — |
| SealQA (12 distractors, >24k) | 65.1 | — |
| SciCode | — | 39.4 |
| SealQA (no distractors, <24k) | 80.7 | — |
| Covered | 49/63 | 21/63 |
Scorecard
Overall 0–100 score — category strengths (50%), shared-benchmark wins (30%), coverage (20%)
| Model | Overall score | Shared benchmark wins | Coverage | Strongest at | Weakest at |
|---|---|---|---|---|---|
| Ripple-65B Preview Winner | 76.4 | 5 of 7 | 21 of 63 (33.0%) | Long Context | General Agents |
| Kolibri-1 | 67.5 | 2 of 7 | 49 of 63 (78.0%) | General Agents | Agentic Coding |
Category strengths
Where each model stands, per benchmark category — score relative to the best model, adjusted for benchmark coverage. This also ranks models when they share few benchmarks.
| Category | Ripple-65B Preview | Kolibri-1 |
|---|---|---|
| STEM & Reasoning 13 benchmarks | 99.6 | 98.6 |
| Agentic Coding 6 benchmarks | 100 | 60.4 |
| General Agents 5 benchmarks | 77.3 | 100 |
| Long Context 2 benchmarks | 100 | 91.1 |
Fully comparable benchmarks
Scored by all 2 selected models (7)
| Benchmark | Category | Kolibri-1 | Ripple-65B Preview |
|---|---|---|---|
| GPQA Diamond | stem_reasoning | 84.3 | 85.2 |
| AA-LCR | long_context | 68.3 | 75 |
| AIME 26 | stem_reasoning | 96 | 95 |
| SWE-bench Verified | coding_agent | 66.4 | 76.5 |
| Terminal Bench 2.1 | coding_agent | 27.7 | 80.5 |
| AIME 2025 | stem_reasoning | 96.9 | 98.3 |
| TauBench V3 Banking | general_agent | 38.1 | 31.27 |
Partial coverage benchmarks
Scored by at least one but not all selected models (56) — missing scores show as —
| Benchmark | Category | Kolibri-1 | Ripple-65B Preview |
|---|---|---|---|
| OmniScience Accuracy 1/2 models | knowledge | 14.8 | — |
| IFBench (prompt loose) 1/2 models | instruction_following | 78.1 | — |
| SealQA (12 distractors, <24k) 1/2 models | agentic | 61 | — |
| LiveCodeBench v6 1/2 models | stem_reasoning | 85.9 | — |
| SWE-bench Pro 1/2 models | coding_agent | — | 61.42 |
| BFCL-V4 1/2 models | general_agent | 61.4 | — |
| FRAMES (<24k) 1/2 models | agentic | 71.2 | — |
| German Public Sector 1/2 models | agentic | 75 | — |
| OmniScience Non-Hallucination 1/2 models | knowledge | 44 | — |
| Aerospace 1/2 models | agentic | 58.9 | — |
| Agentic Wiki QA (DE) 1/2 models | agentic | 69.4 | — |
| AIME 2025 (DE) 1/2 models | stem_reasoning | 87.5 | — |
| AutomationBench (Public) 1/2 models | general_agent | — | 36.4 |
| Overall (EN) 1/2 models | stem_reasoning | 75.5 | — |
| StrongREJECT v2 (adaptive) 1/2 models | safety | — | 97.5 |
| MuSiQue (EN) 1/2 models | agentic | 77.3 | — |
| SWE-bench Pro v2.0 Hard 1/2 models | coding_agent | — | 79.1 |
| SealQA (no distractors, >24k) 1/2 models | agentic | 100 | — |
| HumanEval+ 1/2 models | code | 92.7 | — |
| MMLU-ProX 1/2 models | multilingual | 75.5 | — |
| MMLU-Pro 1/2 models | knowledge | — | 84.2 |
| AIME 2024 1/2 models | stem_reasoning | — | 97.1 |
| BrowseComp 1/2 models | general_agent | 29.4 | — |
| BFCL v4 (web search) 1/2 models | agentic | 62.5 | — |
| RGB Closed-Book 1/2 models | grounding_hallucination | 51 | — |
| Industrial Drive Technology 1/2 models | agentic | 60 | — |
| AA-Omniscience Index (public set) 1/2 models | knowledge | -32.8 | — |
| MMLU 1/2 models | knowledge | — | 89.6 |
| Humanity's Last Exam (DE) 1/2 models | stem_reasoning | 15.9 | — |
| IFBench 1/2 models | instruction_following | — | 79.2 |
| Humanity's Last Exam 1/2 models | stem_reasoning | 21.5 | — |
| BFCL v4 (non-live AST) 1/2 models | agentic | 79.1 | — |
| BFCL v4 (live) 1/2 models | agentic | 78.9 | — |
| BFCL v4 (memory) 1/2 models | agentic | 62.8 | — |
| AIME 2026 (DE) 1/2 models | stem_reasoning | 90 | — |
| Terminal-Bench 2.1 (Terminus-2) 1/2 models | coding_agent | — | 80.5 |
| Automotive Supplier 1/2 models | agentic | 99 | — |
| GPQA Diamond (DE) 1/2 models | stem_reasoning | 81.3 | — |
| HMMT Feb 26 1/2 models | stem_reasoning | — | 90.5 |
| SQuAD (Utility Accuracy) 1/2 models | grounding_hallucination | 83.4 | — |
| BFCL v4 (multi-turn) 1/2 models | agentic | 47.5 | — |
| RGB Fact-Check (Error Correction) 1/2 models | grounding_hallucination | 34 | — |
| SQuAD (M/A Grounding Score) 1/2 models | grounding_hallucination | 23.4 | — |
| LongBenchPro 1/2 models | long_context | 64.5 | — |
| FRAMES (>24k) 1/2 models | agentic | 73 | — |
| Semiconductors 1/2 models | agentic | 80.4 | — |
| Honeypot 1/2 models | agentic | 80.8 | — |
| MCP-Atlas 1/2 models | general_agent | — | 71.67 |
| DeepSWE 1.1 1/2 models | coding_agent | — | 44.3 |
| FORTRESS Benign 1/2 models | safety | — | 95.5 |
| RGB Negative (Abstention) 1/2 models | grounding_hallucination | 85.6 | — |
| MMLU Pro COT 1/2 models | knowledge | 80 | — |
| Overall (DE) 1/2 models | stem_reasoning | 70.8 | — |
| SealQA (12 distractors, >24k) 1/2 models | agentic | 65.1 | — |
| SciCode 1/2 models | — | — | 39.4 |
| SealQA (no distractors, <24k) 1/2 models | agentic | 80.7 | — |
Comparison FAQ
Which model is better, Ripple-65B Preview or Kolibri-1?
Ripple-65B Preview wins this comparison with an overall score of 76 vs 68 for Kolibri-1, leading 5 of 7 shared benchmarks.
Which model is better at STEM & Reasoning, Ripple-65B Preview or Kolibri-1?
Ripple-65B Preview leads the STEM & Reasoning category (strength score 99.6/100 across 3 comparable benchmarks).
Where does Kolibri-1 outperform Ripple-65B Preview?
Kolibri-1 leads the General Agents category (strength score 100/100 across 1 comparable benchmark).
How reliable is this comparison verdict?
Based on 7 of 63 benchmarks scored by all 2 selected models.