Weights & Biases
Coverage
Build and scale faster on the purpose-built AI cloud, engineered from silicon to API.
$10 off the fastest DeepSeek V4.1 Flash, Kimi K3 and GLM 5.3 from Synthetic
Latest news about CoreWeave
Weights & Biases
Official sourceRelease Notes
The W&B documentation site published several new articles in September 2026: a Models guide explaining how to use automations to send a templated prompt to ARIA and start a new conversation on a specified event, a Weave article on creating a custom signal to tag agent turns for a defined behavior, and a Support article Together these represent fresh, technically useful provider updates: ARIA automations extend W&B Models event-driven workflows into LLM app orchestration, the custom-signal feature gives Weave users a way to annotate and analyze agent behavior, and the Weave agent evaluation and caching articles document developer-faci
Weights & Biases
Official sourceRelease Notes
Weights & Biases released W&B Server v0.84.1 on September 2, 2026, shipping targeted bug fixes: the UI now shows all tags within a team up to the 1000-tag maximum instead of truncating at 500, a crash in Reports panel grids triggered by rapid successive interactions was resolved, and tables and artifacts no longer rely These server releases are the most technically substantive provider updates in the candidate set, with concrete version numbers, dates, and developer-facing feature/fix detail directly sourced from docs.wandb.ai. The combination of security/operational fixes (GraphQL introspection, API key lifecycle) and new platform c
About CoreWeave
What they do
Weights & Biases provides MLOps tooling and hosted inference access through its Weave and inference products, letting developers use open models without managing infrastructure.
How they were founded
Founded in 2017 by Lukas Biewald, Chris Van Pelt, and Shawn Lewis as a developer-first machine learning tooling company.
Quick Info
- Organization
- Weights and Biases, LLC
- Headquarters
- San Francisco, California, United States
- SDK package
- @ai-sdk/openai-compatible
- Synced at
- Sep 18, 2026
Models served
20 models available through CoreWeave
These model results are sorted newest first so you can quickly see the latest options from this provider.
CoreWeavedeepseek
DeepSeek V4.1 Flash
deepseek-ai/DeepSeek-V4.1-FlashCoreWeaveReleased Sep 10, 20261,048,576 token context windowIn $0.20 · Out $0.65
Input
Output
CoreWeaveglm
GLM 5.3 Flash
zai-org/GLM-5.3-FlashCoreWeaveReleased Aug 26, 20261,048,576 token context windowIn $0.15 · Out $0.50
Input
Output
CoreWeavegranite
Granite 4.2 8B
ibm-granite/granite-4.2-8bCoreWeaveReleased Aug 24, 2026131,072 token context windowIn $0.10 · Out $0.15
Input
Output
CoreWeaveqwen
Input
Output
CoreWeavedeepseek
DeepSeek V4 Pro 0813
deepseek-ai/DeepSeek-V4-Pro-0813CoreWeaveReleased Aug 13, 20261,048,576 token context windowIn $1.31 · Out $3.96
Input
Output
CoreWeavenemotron
Nemotron 3.5 Lightning
nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3BCoreWeaveReleased Aug 11, 2026262,144 token context windowIn $0.07 · Out $0.20
Input
Output
CoreWeavedeepseek
DeepSeek V4 Flash 0731
deepseek-ai/DeepSeek-V4-Flash-0731CoreWeaveReleased Jul 31, 2026262,144 token context windowIn $0.13 · Out $0.28
Input
Output
CoreWeaveglm
Input
Output
CoreWeaveminimax-m3
Input
Output
CoreWeavekimi-k2
Kimi K2.7 Code
moonshotai/Kimi-K2.7-CodeCoreWeaveReleased Jun 12, 2026262,144 token context windowIn $0.71 · Out $3.50
Input
Output
CoreWeavenemotron
Nemotron 3 Ultra
nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55BCoreWeaveReleased Jun 4, 2026262,144 token context windowIn $0.50 · Out $2.15
Input
Output
CoreWeavekimi-k2
Input
Output
CoreWeaveqwen3.6
Input
Output
CoreWeavegemma
Gemma 4 26B A4B
google/gemma-4-26B-A4B-itCoreWeaveReleased Apr 2, 2026262,144 token context windowIn $0.10 · Out $0.30
Input
Output
CoreWeavegemma
Input
Output
CoreWeavedeepseek
DeepSeek V3.1
deepseek-ai/DeepSeek-V3.1CoreWeaveReleased Aug 21, 2025161,000 token context windowIn $0.55 · Out $1.65
Input
Output
CoreWeavegpt-oss
Input
Output
CoreWeavegpt-oss
Input
Output