Thinkymachines releases Inkling-Small: 276B-parameter open-source multimodal model with 140 tps inference at $11
_akhaliq · x · 2026-07-31
Thinkymachines released Inkling-Small, an open-source multimodal model with 276B total parameters (12B active), supporting text, image, and audio. It achieves 140 tokens per second on inference endpoints at $11, with performance comparable to larger models. Full weights are available.
Related event: Thinking Machines Unveils Inkling-Small: 276B MoE Model with 1M Context(32 posts)→
More from Models
- Open-Weight Inkling Small Tops ARC-AGI at $0.23/Task — LiTianleli · 2026-07-31
- AI Model Boundary Pushing: Anthropic's Model Exhibits Edgy Conversational Style — repligate · 2026-07-31
- Frequent LLM Price Drops Break Static Routing, Strengthening Case for AI Gateways — shensi · 2026-07-31
- Luna Model Matches Heavyweights in Web Agent Benchmarks at 1/20th the Cost — kohjingyu · 2026-07-31
- Gemini Image Recognition Seems Broken, Spitting Random Guesses — sazzer22 · 2026-07-31
- User Reports Claude Opus Unintentionally Leaking Its Own Jailbreak Prompts — Kyrannio · 2026-07-31