Compare / Ripple 65B Preview vs Llama 3.3 70B
Ripple 65B Preview vs Llama 3.3 70B
Updated November 2026
Verdict
Ripple 65B Preview scores higher on the Small LM Index: 74.6 vs 33.8. On the 5 benchmarks both labs report, Ripple 65B Preview wins 5 and Llama 3.3 70B wins 0. Ripple 65B Preview is smaller: it needs about 41.9 GB of memory at 4-bit, against 47.4 GB for Llama 3.3 70B. Only Llama 3.3 70B has a hosted API on OpenRouter ($0.15 per million tokens); the other has to be self-hosted.
Pick Ripple 65B Preview if…
- you want the strongest model overall
- you need math and reasoning
- you need precise instruction following
- you have limited memory (about 41.9 GB at 4-bit)
- you work with long documents
- you need a permissive license (Apache-2.0)
Pick Llama 3.3 70B if…
- you want a hosted API rather than running it yourself
- you want the more widely used model (more quantized builds and community support)
Specifications
| Ripple 65B Preview | Llama 3.3 70B | |
|---|---|---|
| Small LM Index | 74.6 (likely 74–77) | 33.8 (likely 31–37) |
| Lab | Ricursive Intelligence · Other | Meta · United States |
| Released | 2026-11-04 | 2024-12-06 |
| Parameters | 64.8B (dense) | 70.6B (dense) |
| Memory at 4-bit | ~41.9 GB | ~47.4 GB |
| Context window | 1M tokens | 128K tokens |
| Input | text | text |
| Reasoning | Always thinks | No thinking |
| License | Apache-2.0 | Llama 3.3 Community |
| API price / 1M tokens | Self-host only | $0.15 ($0.10 in, $0.32 out) |
| Hugging Face downloads (30 days) | 15K | 344.1K |
Benchmarks
Higher is better. Bold marks the winner where both labs report a score. ≈ marks the index model's estimate for a score that wasn't reported, ✓ an independent measurement and * a score measured by another lab.
| Benchmark | Ripple 65B Preview | Llama 3.3 70B |
|---|---|---|
| MMLU-Pro Knowledge | 84.2 | 71.3✓ |
| GPQA Diamond Knowledge | 85.2 | 49.8✓ |
| Humanity's Last Exam Reasoning | ≈ 18 | 3.6✓ |
| AIME Math | 98.3 | 30.0✓ |
| LiveCodeBench Coding | ≈ 90 | 28.8✓ |
| SWE-bench Verified Coding | 76.5 | ≈ 36 |
| Terminal-Bench Agentic | 80.5 | — |
| τ²-Bench Agentic | — | 26.6✓ |
| IFBench Instruction following | 79.2 | 47.1✓ |
| MMLU Knowledge | 89.6 | — |
| MATH-500 Math | — | 77.3✓ |
| HMMT Math | 90.5 | — |
| AA-LCR Long context | 75.0 | 15.7✓ |
| SciCode Coding | 39.4 | — |
Scores are mostly reported by each lab with its own settings, so small gaps are noise. See the methodology. Open in the interactive comparison to add more models.
Questions
Which is better, Ripple 65B Preview or Llama 3.3 70B?
Ripple 65B Preview scores higher on the Small LM Index: 74.6 vs 33.8. On the 5 benchmarks both labs report, Ripple 65B Preview wins 5 and Llama 3.3 70B wins 0.
Which is cheaper to run?
Only Llama 3.3 70B has a hosted API on OpenRouter ($0.15 per million tokens); the other has to be self-hosted. Ripple 65B Preview is smaller: it needs about 41.9 GB of memory at 4-bit, against 47.4 GB for Llama 3.3 70B.
Which has the longer context window?
Ripple 65B Preview has the longer context window: 1M tokens against 128K.
Can I use Ripple 65B Preview and Llama 3.3 70B commercially?
Ripple 65B Preview is released under Apache-2.0 and Llama 3.3 70B under Llama 3.3 Community. Apache 2.0 and MIT allow commercial use; other licenses can add conditions, so read the license before shipping a product.