TPSTokens Per Second
model release

Ollama v0.40.16 adds Ricursive Intelligence's 65B dense Ripple Preview coding model

TL;DR

Ollama v0.40.16 adds Ripple Preview from Ricursive Intelligence, a 65B dense model built for agentic coding, runnable with `ollama run ripple-preview`. The release also cuts memory use for Qwen3.8 Flash Next on Apple silicon by about 27 GiB.

1 min read
0

Ollama v0.40.16 adds Ripple Preview, a 65B dense model from Ricursive Intelligence built for agentic coding. Users can run it with ollama run ripple-preview. The release was published on GitHub on 09 Nov by github-actions.

Ripple Preview

According to the release notes, Ripple Preview is a 65B dense model built for agentic coding. Context window, pricing and benchmark scores are not yet disclosed.

Other changes in v0.40.16

  • Qwen3.8 Flash Next uses about 27 GiB less memory on Apple silicon.
  • ollama launch claude always routes Claude Code to the selected model.
  • ollama launch explains when a cloud model has been retired.
  • /v1/responses respects truncation and no longer silently truncates by default.
  • /v1/models returns an empty list instead of null when no models are installed.
  • CUDA 13.5 support for Windows on ARM.
  • Updated llama.cpp.

The full changelog is listed on GitHub as v0.40.15...v0.40.16.

What this means

Developers running local or self-hosted coding agents can now pull a 65B dense coding model through Ollama with a single command. The /v1/responses change matters for anyone building on Ollama's OpenAI-compatible endpoints: requests are no longer silently truncated by default, so truncation behavior follows what the request specifies. The /v1/models fix, which returns an empty list rather than null on a machine with no models installed, removes a case that client code would otherwise have to handle separately.

Source: github.com ↗
OllamaRipple PreviewRicursive Intelligenceagentic codinglocal modelsQwen3.8 Flash Nextllama.cpp

Related Articles

model release

Anthropic releases Claude Fable 5.2 and Mythos 5.2 at $8/$40 per 1M tokens, with an optional fast mode

Anthropic released Claude Fable 5.2 and Claude Mythos 5.2, priced at $8 per million input tokens and $40 per million output tokens. Anthropic says they generate output up to 30% faster than the 5.1 models, and Fable 5.2 reaches 49.1% on FrontierCode 2.0 at Xhigh effort.

model release

Bytedance Seedream 5.5 claims pixel-stable image editing across up to 100 rounds

Bytedance Seed has introduced Seedream 5.5, an image generation model for professional production workflows. According to Bytedance, it keeps unchanged areas pixel-identical across up to 100 editing rounds and generates native RGBA images with transparent backgrounds.

model release

Meta releases Muse Ember, its lightest model, at 10 cents per million input tokens

Meta Superintelligence Lab released Muse Ember on Monday, its lightest and most cost-efficient model yet, priced at 10 cents per million input tokens. Meta says it is built for repetitive agent work and pairs with the larger Muse Wonder.

model release

Google releases Gemini 4 Argon for robot control, reports 87% on ERQA without helper tools

Google has released Gemini 4 Argon, a general-purpose physical-reasoning model for robot control, Real2Sim scene construction, robot video labeling and iterative policy improvement. Google reports 87% on the ERQA benchmark using pure reasoning, and says that leads other frontier models. It is available to Google AI Ultra subscribers and via the paid Gemini API as gemini-4-argon.

Comments

Loading...