Thinking Machines' Inkling 2 Small scores 31 on AA Intelligence Index, planned weight release would top non-China open models
ArtificialAnlys · x · 2026-11-06
Thinking Machines released Inkling 2 Small, scoring 31 on the Artificial Analysis Intelligence Index v5.0.1 — up 18 points from Inkling Small. If weights are released as planned (Apache 2.0), it would be the highest-scoring open-weights model developed outside China.
Key facts
- 594B total / 17B active parameters, text+image input; launches proprietary via Tinker and Tinker Playground, weights to follow
- Pricing: $0.30/$1.00 per 1M input/output tokens ($0.03 cached); $1.18 per Intelligence Index task (6x Inkling Small) but far cheaper than Gemini 3.8 Flash (high, $7.77) or Claude Sonnet 5.5 (medium, $10.02)
- Big agentic gains: AutomationBench-AA 5%→46%, FrontierCode 2.0 5%→25%, HLE-Diamond 12%→29%
- Complementary coding profile vs Claude Sonnet 5.5: 25% vs 11% on FrontierCode 2.0, but 13% vs 30% on Terminal-Bench 4.0
- Hallucination rate fell 61%→27%, mostly from answering less (attempt rate 77%→53%); accuracy only 32%→34%
- AA-Briefcase long-horizon workflows: 1527 Elo, above Claude Sonnet 5.5 (1439) and Gemini 3.8 Flash (1203)
- AA-3D 1021 and Code2Video-AA 1020 Elo, near Muse Spark 1.3 and Qwen3.8 Max
- Heavy reasoner: 150K output tokens per task, 5x Claude Sonnet 5.5
More from Models
- Inkling 2 Small scores 31 on AA index; finetunability may become a key open-weight metric — simonguozirui · 2026-11-06
- Small-model customizability called critical as Inkling 2 Small touts SOTA fine-tuning — clarejtbirch · 2026-11-06
- TensorFold 1.2.3 boosts Flash Next decode 21–23%; Apache-2.0 Ripple-65B Preview lands — AIFlow_ML · 2026-11-06
- Researcher alth0u teases that a new model's training run is underway — clarejtbirch · 2026-11-06
- Inkling 2 Small pretrained on 8.4T tokens in 68 hours, natively in NVFP4 — clarejtbirch · 2026-11-06
- Why do Astra and Sol 6.1 perform so well? Multi-agent RL and epistemic vigilance — scaling01 · 2026-11-06