Step 5 Preview is StepFun's flagship model for agentic work, built on a sparse Mixture-of-Experts architecture (27B active / 600B total parameters). It performs strongly in software engineering and professional...
Published: 08/10/2026
https://openrouter.ai/stepfun/step-5-preview
Claude Haiku 5.5 is Anthropic's small, fast model for high-volume, cost-sensitive work such as summarization, subagents, and browser use. It succeeds Claude Haiku 4.5 with stronger coding, computer use, and...
Published: 07/10/2026
https://openrouter.ai/anthropic/claude-haiku-5.5
Nano Banana 2.1 (Gemini Nano Banana 2.1) is Google's image generation and editing model on the Flash tier, succeeding Nano Banana 2 and Nano Banana Pro. It improves product recontextualization,...
Published: 06/10/2026
https://openrouter.ai/google/gemini-nano-banana-2.1
<!-- openclaw-release-publication:docs-v1 -->
64 direct commits · 4,179 pull requests · 503 contributors
The release notes and changelog contain the same content, presented in two formats:
@0xalydev @0xghost42 @100yenadmin @1052 @373246784-design @609nft @849261680 @aaajiao @aadi-joshi @aanalog-crow @aaronqianaa @aaronwander @aatreya @abacha @abhishek17n @adamamzalag @adele-with-a-b @adembektas @agebreak0083 @akagifreeez @akarshghale @aleps001 @alex-govorov @alexph-dev @alexscfraser @alix-007 @altaywtf @alvinveroy @amadeubordaneto @amittell @amknight @amtls-bbq @andersonjeccel @angeliti999 @aniruddhaadak80 @ansxor @anusha0501 @anyech @arnehorn @aron-intframe @a...
Published: today https://github.com/openclaw/openclaw/releases/tag/v2026.9.5
Latest published Linux companion: v2026.9.5
Published: today https://github.com/openclaw/openclaw/releases/tag/linux-stable
Bonsai 2 27B is a 27B-parameter reasoning model from PrismML derived from Qwen3.8-27B. It supports coding, mathematics, tool calling, and image understanding with a 262K-token context window. Ternary compression shrinks...
Published: Fri, 18 Sep 2026 17:54:06 GMT
GLM-5.3-FlashX is the high-speed variant of Z.ai's GLM-5.3-Flash, a native multimodal model delivering inference speeds of up to 200 tokens/s. Built on the same hybrid sparse and linear attention architecture...
Published: Fri, 18 Sep 2026 15:07:00 GMT
Hoping to tackle the 50% of textile emissions associated with the dyeing process, a vaunted French startup has created non-toxic dyes that require substantially less water to activate and last just as long. Additionally, they take hold electromagnetically at room temperature, rather than over long periods at boiling temperatures, and don’t require chemical fixatives. EverDye’s […] The post Company’s Textile Dyes Work Through Magnetism, Cutting Fashion’s Water-Use, E...
Why this is uplifting: This story highlights positive developments and offers reasons for hope and optimism.
Published: Fri, 18 Sep 2026 18:30:36 +0000