Model tracker / Ricursive Intelligence
Ripple-65B Preview
Dense 65B open-weight model for agentic coding and reasoning, text only
Ripple-65B Preview is a fast open-weight model from Ricursive Intelligence, released 3 November 2026. It's in preview. At Together AI it costs $0.20 per million input tokens and $0.60 per million output tokens, with a 131K-token context window.
- Made by
- Ricursive Intelligence
- Released
- 2026-11-03
- Status
- Preview, new
- Tier
- Fast
- Weights
- Open
- Memory to run
- About 156 GB (BF16)
- Context window
- 131,072 tokens
- Input
- text
- API id
ricursive-ai/Ripple-65B-Preview
Ripple-65B Preview price at Together AI
US dollars per million tokens. Blended is 3 input tokens for every output token, one number to compare on.
| Provider | Input | Output | Cached input | Blended | Batch | Checked |
|---|---|---|---|---|---|---|
| Together AI | $0.20 | $0.60 | $0.02 | $0.30 | – | 2026-11-06 |
- Together AI: Together lists a 128K context window for its serverless endpoint; the model supports up to 1M.
Ripple-65B Preview benchmarks
| Benchmark | Score | What it measures | Reported by |
|---|---|---|---|
| SWE-bench Verified | 76.5% | real GitHub issues fixed | Ricursive Intelligence |
| GPQA Diamond | 85.2% | graduate-level science questions answered correctly | Ricursive Intelligence |
Scores reported by a model's maker aren't independently checked.
Can you run Ripple-65B Preview yourself?
Yes. Ricursive Intelligence publishes the weights on Hugging Face as ricursiveAI/Ripple-65B-Preview, under Apache 2.0: 64.8B parameters, 130 GB in BF16.
Running it takes about 156 GB of GPU memory: a server with several data-centre GPUs (eight 80 GB GPUs hold 640 GB). That's the weight files plus 20% for working memory with short inputs; long contexts need more, and fitting in memory doesn't make it fast.
Ricursive Intelligence also publishes a smaller NVFP4 version, ricursiveAI/Ripple-65B-Preview-NVFP4, at 57 GB: about 69 GB to run, so one 80 GB data-centre GPU, or a Mac with comfortably more unified memory than that. NVFP4 needs NVIDIA's Blackwell-generation GPUs.
Models like Ripple-65B Preview
Other fast models closest in price, at each maker's own price or the cheapest provider.
| Model | Input | Output | Context | Weights |
|---|---|---|---|---|
| Ripple-65B Preview | $0.20 | $0.60 | 131K | Open |
| Gemma 4 26B A4B Google | $0.15 | $0.60 | 262K | Open |
| MiniMax M3.1 Flash MiniMax | $0.15 | $0.60 | 1M | Open |
| Mistral Small 4 Mistral AI | $0.15 | $0.60 | 256K | Open |
| GLM-5.3 Flash Z.ai | $0.15 | $0.50 | 1M | Open |
| NVIDIA Nemotron 3.5 Super VL NVIDIA | $0.15 | $0.50 | 262K | Open |
Compare them all in the model tracker →
Ripple-65B Preview on dbggr
- Update apt now: a Debian and Ubuntu bug lets a network attacker swap packages and run code as root Daily, 2026-11-04
Ripple-65B Preview FAQ
How much does Ripple-65B Preview cost?
Ripple-65B Preview costs $0.20 per million input tokens and $0.60 per million output tokens at Together AI, and $0.02 per million cached input tokens. Prices are list prices in US dollars, last checked 6 November 2026.
What is Ripple-65B Preview's context window?
Ripple-65B Preview takes up to 131,072 tokens in one request.
Is Ripple-65B Preview open weight?
It's listed as open weight, but it's still in preview: check Ricursive Intelligence's announcement for when the weights and licence are published.
Can I run Ripple-65B Preview locally?
Ripple-65B Preview's weights are 130 GB in BF16, so it needs about 156 GB of GPU memory: a server with several data-centre GPUs (eight 80 GB GPUs hold 640 GB). Ricursive Intelligence also publishes a NVFP4 version that needs about 69 GB.
Where can I use Ripple-65B Preview?
Ripple-65B Preview is available from Together AI. At Together AI, call it as ricursive-ai/Ripple-65B-Preview.