Ripple-65B Preview
Explore Ripple-65B Preview from Ricursive Intelligence: published specifications, source-linked vendor benchmarks, independent evaluator coverage and recorded pricing when available.
Compare Ripple-65B Preview with other models →Compare subscription plansExplore data coverage
Pricing · OpenRouter
Dated cached OpenRouter rates in USD per 1M tokens. Open the dashboard for live enhancements. Per-metric endpoint minima can refer to different providers; they are not a guaranteed combined rate from one endpoint.
| Tier | Input / 1M | Output / 1M | Cached input / 1M | Cache write / 1M | Date & source |
|---|---|---|---|---|---|
| Default | $0.1 | $0.3 | $0.02 | — | 2026-11-10 · OpenRouter source |
Recorded pricing notes
Single GMICloud endpoint ricursiveai/ripple-65b-preview-20261104 (bf16, 16,384-token context); at the 5 November check its status was -5 with 0% uptime over 5 minutes and about 21% over a day, so no healthy price is shown. No contributor sibling at this check; min-healthy endpoint minima 2026-11-10
Official / vendor benchmarks
Default headline records. Own-vendor, peer-vendor and third-party provenance remain visible in evidence; configurations may differ.
| Benchmark / evaluator | Headline score | Evidence |
|---|---|---|
| Terminal-Bench 2.1 | 80.5%unknown effort | The recorded result & source80.5% · raw 80.52 % Headline · unknown effort · Own vendor Source/record date: 2026-11-04 Ricursive Ripple-65B Preview Hugging Face card (revision 240d5966cbe0), instruct table: rows “Terminal-Bench 2.1 (self-reported)” and “Terminal-Bench 2.1 (Terminus 2)” both print 80.5; the card and release text state 80.52, used here. The release blog tooltip labels the run “RL step 700, Terminus-2”. Harness: Terminus 2. Ricursive: all models tested at full BF16 precision with maximum effort levels; the Ripple chat template has no effort setting, so effort is unknown. https://huggingface.co/ricursiveAI/Ripple-65B-Preview/blob/240d5966cbe00f0627c699e59ef58560b60303cf/README.md |
| MMLU | 89.6%unknown effort | The recorded result & source89.6% · raw 89.6 % Headline · unknown effort · Own vendor Source/record date: 2026-11-04 Ricursive Ripple-65B Preview Hugging Face card (revision 240d5966cbe0), instruct table, row “MMLU”. Ricursive: all models tested at full BF16 precision with maximum effort levels; the Ripple chat template has no effort setting, so effort is unknown. https://huggingface.co/ricursiveAI/Ripple-65B-Preview/blob/240d5966cbe00f0627c699e59ef58560b60303cf/README.md |
| MMLU-Pro | 84.2%unknown effort | The recorded result & source84.2% · raw 84.2 % Headline · unknown effort · Own vendor Source/record date: 2026-11-04 Ricursive Ripple-65B Preview Hugging Face card (revision 240d5966cbe0), instruct table, row “MMLU-Pro”. Ricursive: all models tested at full BF16 precision with maximum effort levels; the Ripple chat template has no effort setting, so effort is unknown. https://huggingface.co/ricursiveAI/Ripple-65B-Preview/blob/240d5966cbe00f0627c699e59ef58560b60303cf/README.md |
| GPQA Diamond | 85.2%unknown effort | The recorded result & source85.2% · raw 85.2 % Headline · unknown effort · Own vendor Source/record date: 2026-11-04 Ricursive Ripple-65B Preview Hugging Face card (revision 240d5966cbe0), instruct table, row “GPQA Diamond”. Ricursive: all models tested at full BF16 precision with maximum effort levels; the Ripple chat template has no effort setting, so effort is unknown. https://huggingface.co/ricursiveAI/Ripple-65B-Preview/blob/240d5966cbe00f0627c699e59ef58560b60303cf/README.md |
| AIME 2024 | 97.1%unknown effort | The recorded result & source97.1% · raw 97.1 % Headline · unknown effort · Own vendor Source/record date: 2026-11-04 Ricursive Ripple-65B Preview Hugging Face card (revision 240d5966cbe0), instruct table, row “AIME 2024”. Ricursive: all models tested at full BF16 precision with maximum effort levels; the Ripple chat template has no effort setting, so effort is unknown. https://huggingface.co/ricursiveAI/Ripple-65B-Preview/blob/240d5966cbe00f0627c699e59ef58560b60303cf/README.md |
| AIME 2025 | 98.3%unknown effort | The recorded result & source98.3% · raw 98.3 % Headline · unknown effort · Own vendor Source/record date: 2026-11-04 Ricursive Ripple-65B Preview Hugging Face card (revision 240d5966cbe0), instruct table, row “AIME 2025”. Ricursive: all models tested at full BF16 precision with maximum effort levels; the Ripple chat template has no effort setting, so effort is unknown. https://huggingface.co/ricursiveAI/Ripple-65B-Preview/blob/240d5966cbe00f0627c699e59ef58560b60303cf/README.md |
| AIME 2026 | 95%unknown effort | The recorded result & source95% · raw 95 % Headline · unknown effort · Own vendor Source/record date: 2026-11-04 Ricursive Ripple-65B Preview Hugging Face card (revision 240d5966cbe0), instruct table, row “AIME 2026”. Ricursive: all models tested at full BF16 precision with maximum effort levels; the Ripple chat template has no effort setting, so effort is unknown. https://huggingface.co/ricursiveAI/Ripple-65B-Preview/blob/240d5966cbe00f0627c699e59ef58560b60303cf/README.md |
| HMMT February 2026 | 90.5%unknown effort | The recorded result & source90.5% · raw 90.5 % Headline · unknown effort · Own vendor Source/record date: 2026-11-04 Ricursive Ripple-65B Preview Hugging Face card (revision 240d5966cbe0), instruct table, row “HMMT Feb 2026”. Ricursive: all models tested at full BF16 precision with maximum effort levels; the Ripple chat template has no effort setting, so effort is unknown. https://huggingface.co/ricursiveAI/Ripple-65B-Preview/blob/240d5966cbe00f0627c699e59ef58560b60303cf/README.md |
| SWE-bench Verified | 76.5%unknown effort | The recorded result & source76.5% · raw 76.5 % Headline · unknown effort · Own vendor Source/record date: 2026-11-04 Ricursive Ripple-65B Preview Hugging Face card (revision 240d5966cbe0), instruct table, row “SWE-bench Verified”. Harness: OpenCode (the card says Ricursive uses OpenCode wherever no harness is named). Ricursive: all models tested at full BF16 precision with maximum effort levels; the Ripple chat template has no effort setting, so effort is unknown. https://huggingface.co/ricursiveAI/Ripple-65B-Preview/blob/240d5966cbe00f0627c699e59ef58560b60303cf/README.md |
| SWE-bench Pro v1 | 61.4%unknown effort | The recorded result & source61.4% · raw 61.42 % Headline · unknown effort · Own vendor Source/record date: 2026-11-04 Ricursive Ripple-65B Preview Hugging Face card (revision 240d5966cbe0), instruct table, row “SWE-bench Pro v1.0”. Harness: OpenCode (the card says Ricursive uses OpenCode wherever no harness is named). Ricursive: all models tested at full BF16 precision with maximum effort levels; the Ripple chat template has no effort setting, so effort is unknown. https://huggingface.co/ricursiveAI/Ripple-65B-Preview/blob/240d5966cbe00f0627c699e59ef58560b60303cf/README.md |
| SWE-bench Pro v2-Hard | 79.1%unknown effort | The recorded result & source79.1% · raw 79.1 % Headline · unknown effort · Own vendor Source/record date: 2026-11-04 Ricursive Ripple-65B Preview Hugging Face card (revision 240d5966cbe0), instruct table, row “SWE-bench Pro v2.0 Hard”. Harness: Terminus 2 (stated in the card for Ripple-65B Preview). Ricursive: all models tested at full BF16 precision with maximum effort levels; the Ripple chat template has no effort setting, so effort is unknown. https://huggingface.co/ricursiveAI/Ripple-65B-Preview/blob/240d5966cbe00f0627c699e59ef58560b60303cf/README.md |
| DeepSWE v1.1 | 44.3%unknown effort | The recorded result & source44.3% · raw 44.3 % Headline · unknown effort · Own vendor Source/record date: 2026-11-04 Ricursive Ripple-65B Preview Hugging Face card (revision 240d5966cbe0), instruct table, row “DeepSWE 1.1”. Harness: Terminus 2 (stated in the card for Ripple-65B Preview). Ricursive: all models tested at full BF16 precision with maximum effort levels; the Ripple chat template has no effort setting, so effort is unknown. https://huggingface.co/ricursiveAI/Ripple-65B-Preview/blob/240d5966cbe00f0627c699e59ef58560b60303cf/README.md |
| SciCode (Ricursive, Inspect AI; aggregation unspecified) | 39.4%unknown effort | The recorded result & source39.4% · raw 39.4 % Headline · unknown effort · Own vendor Source/record date: 2026-11-04 Ricursive Ripple-65B Preview Hugging Face card (revision 240d5966cbe0), instruct table, row “SciCode (Inspect AI)”. Harness: Inspect AI; aggregation (subtask or problem) unstated. Ricursive: all models tested at full BF16 precision with maximum effort levels; the Ripple chat template has no effort setting, so effort is unknown. https://huggingface.co/ricursiveAI/Ripple-65B-Preview/blob/240d5966cbe00f0627c699e59ef58560b60303cf/README.md |
| τ³-bench Banking | 31.3%unknown effort | The recorded result & source31.3% · raw 31.27 % Headline · unknown effort · Own vendor Source/record date: 2026-11-04 Ricursive Ripple-65B Preview Hugging Face card (revision 240d5966cbe0), instruct table, row “τ³-bench banking”. Harness: OpenCode (the card says Ricursive uses OpenCode wherever no harness is named). Ricursive: all models tested at full BF16 precision with maximum effort levels; the Ripple chat template has no effort setting, so effort is unknown. https://huggingface.co/ricursiveAI/Ripple-65B-Preview/blob/240d5966cbe00f0627c699e59ef58560b60303cf/README.md |
| AutomationBench (public split) | 36.4%unknown effort | The recorded result & source36.4% · raw 36.4 % Headline · unknown effort · Own vendor Source/record date: 2026-11-04 Ricursive Ripple-65B Preview Hugging Face card (revision 240d5966cbe0), instruct table, row “AutomationBench (Public)”. Harness: OpenCode (the card says Ricursive uses OpenCode wherever no harness is named). Ricursive: all models tested at full BF16 precision with maximum effort levels; the Ripple chat template has no effort setting, so effort is unknown. https://huggingface.co/ricursiveAI/Ripple-65B-Preview/blob/240d5966cbe00f0627c699e59ef58560b60303cf/README.md |
| MCP Atlas | 71.7%unknown effort | The recorded result & source71.7% · raw 71.67 % Headline · unknown effort · Own vendor Source/record date: 2026-11-04 Ricursive Ripple-65B Preview Hugging Face card (revision 240d5966cbe0), instruct table, row “MCP Atlas”. Harness: OpenCode (the card says Ricursive uses OpenCode wherever no harness is named). Ricursive: all models tested at full BF16 precision with maximum effort levels; the Ripple chat template has no effort setting, so effort is unknown. https://huggingface.co/ricursiveAI/Ripple-65B-Preview/blob/240d5966cbe00f0627c699e59ef58560b60303cf/README.md |
| FORTRESS Adversarial (Ricursive; higher is better) | 83.8%unknown effort | The recorded result & source83.8% · raw 83.8 % Headline · unknown effort · Own vendor Source/record date: 2026-11-04 Ricursive Ripple-65B Preview Hugging Face card (revision 240d5966cbe0), instruct table, row “FORTRESS Adversarial”. Higher is better; split, judge and formula unpublished. Ricursive: all models tested at full BF16 precision with maximum effort levels; the Ripple chat template has no effort setting, so effort is unknown. https://huggingface.co/ricursiveAI/Ripple-65B-Preview/blob/240d5966cbe00f0627c699e59ef58560b60303cf/README.md |
| FORTRESS Benign (Ricursive; higher is better) | 95.5%unknown effort | The recorded result & source95.5% · raw 95.5 % Headline · unknown effort · Own vendor Source/record date: 2026-11-04 Ricursive Ripple-65B Preview Hugging Face card (revision 240d5966cbe0), instruct table, row “FORTRESS Benign”. Higher is better; split, judge and formula unpublished. Ricursive: all models tested at full BF16 precision with maximum effort levels; the Ripple chat template has no effort setting, so effort is unknown. https://huggingface.co/ricursiveAI/Ripple-65B-Preview/blob/240d5966cbe00f0627c699e59ef58560b60303cf/README.md |
| StrongREJECT (Ricursive; higher is better) | 97.5%unknown effort | The recorded result & source97.5% · raw 97.5 % Headline · unknown effort · Own vendor Source/record date: 2026-11-04 Ricursive Ripple-65B Preview Hugging Face card (revision 240d5966cbe0), instruct table, row “StrongREJECT”. Higher is better; behaviours, attacks and judge unpublished. Ricursive: all models tested at full BF16 precision with maximum effort levels; the Ripple chat template has no effort setting, so effort is unknown. https://huggingface.co/ricursiveAI/Ripple-65B-Preview/blob/240d5966cbe00f0627c699e59ef58560b60303cf/README.md |
| IFBench | 79.2%unknown effort | The recorded result & source79.2% · raw 79.2 % Headline · unknown effort · Own vendor Source/record date: 2026-11-04 Ricursive Ripple-65B Preview Hugging Face card (revision 240d5966cbe0), instruct table, row “IFBench”. Variant (prompt or instruction level, strict or loose) unstated. Ricursive: all models tested at full BF16 precision with maximum effort levels; the Ripple chat template has no effort setting, so effort is unknown. https://huggingface.co/ricursiveAI/Ripple-65B-Preview/blob/240d5966cbe00f0627c699e59ef58560b60303cf/README.md |
| AA-LCR | 75%unknown effort | The recorded result & source75% · raw 75 % Headline · unknown effort · Own vendor Source/record date: 2026-11-04 Ricursive Ripple-65B Preview Hugging Face card (revision 240d5966cbe0), instruct table, row “AA-LCR”. Version unstated. Ricursive: all models tested at full BF16 precision with maximum effort levels; the Ripple chat template has no effort setting, so effort is unknown. https://huggingface.co/ricursiveAI/Ripple-65B-Preview/blob/240d5966cbe00f0627c699e59ef58560b60303cf/README.md |
Independent evaluators
Evaluator harnesses are distinct from vendor measurements. Missing coverage is not a failed test.
| Benchmark / evaluator | Headline score | Evidence |
|---|---|---|
| DataAnalyticsBench · questions correct /43 · Plotly DataAnalyticsBench | 35 correctunknown effort | The recorded result & source35 correct · raw 35 correct Headline · unknown effort · Independent evaluator Source/record date: 2026-11-07 Harness: Plotly Studio Embedded, one run on 43 private English data-analytics questions in a synthetic wind-energy warehouse. Claude Sonnet 5 judges against Plotly answer key. One or two questions can be noise; no error bars. Exact checkpoint hash, run date and effort unpublished. Company-operated analytics benchmark, distinct from provider results. Table row as displayed: “Ripple-65B Preview† Nov 4, 2026 35 / 43 $2.21 36.4M 2.1 min / 3.2 min 2 h 18 min”. Deep dive of 5 November. † marks open-weight models. Cost is at OpenRouter prices on the run day; the date column is the release date Plotly lists. Serving precision unstated. https://plotly.com/blog/introducing-dataanalyticsbench/ |
Read how we select and source scores or the comparison guide.