Skip to content
Model Pareto

Ripple-65B Preview

Ricursive IntelligenceOpen weightshuggingface.co ↗
Input
—
Output
—
Speed
~76tok/sest.
Context
1Mtokens

Open weights (Apache 2.0), 65B dense Transformer pretrained on 12T tokens; light post-training focused on agentic coding. Card headline: 80.52 on Terminal-Bench 2.1, a version our Terminal-Bench 4.0 does not accept, so it is left out. No API price or measured speed yet.

Rankings

Self-hostingEstimated

How we estimate →

What it would cost to run these open weights yourself on rented GPUs. No API sells this model, so this is its price on the chart.

Hardware
1× B200 180GB
FP8 weights
Throughput
~2,559 tok/s
many requests batched
Price
$0.32–2.84 /M tok
busy → light use
On your own machine
Mac Studio M5 Max 128GB (int4, ~13 tok/s single-stream)
single consumer GPU or Mac
Assumptions (5)
  • FP8 weights (65B params) + 40% KV-cache headroom ≈ 91 GB
  • 1× B200 180GB at $5.98–$14.24/GPU-hour on-demand (2026-09-24)
  • ~2,559 output tok/s aggregate at batch 73 (bandwidth-bound)
  • Blended 3:1 input:output; prefill ~16,356 tok/s
  • Low = 75% utilization at the low GPU price; high = 20% at the high price

Benchmark scores

~ italic, dashed = no published score yet, estimated from related benchmarks.

Coding

SciCodeEstimated
~#20 of 119 (estimated)~57.0%
  1. #1Claude Fable 5.267.7%
  2. #2Claude Opus 5.566.9%
  3. #19Grok 4.757.4%
  4. ~#20Ripple-65B PreviewEstimated57.0%
  5. #20Gemini 3.8 Flash56.6%
No published score — estimated from related benchmarks

Writing & chat

Hard reasoning