

Reasoning-Modelle — Synthszr Charts
Diese Kategorie umfasst KI-Modelle, die auf mehrstufiges logisches Schließen, Mathematik und komplexe Problemlösung ausgelegt sind, statt nur auf schnelle Textgenerierung. Dazu zählen Modelle wie o1, o3, DeepSeek R1 und DeepSeek V3, die vor der Antwort interne Zwischenschritte durchlaufen. Eingesetzt werden sie unter anderem für Programmieraufgaben, wissenschaftliche Fragestellungen und Beweisführung, etwa bei Systemen wie AlphaProof Nexus.
Die Synthszr Charts ranken diese Produkte nach Momentum: Wie oft und wie stark ein Modell aktuell in Tausenden News- und Newsletter-Quellen auftaucht, gewichtet nach Aktualität der Erwähnung. Die Auswertung erfolgt versions-granular, MAI-Thinking-1 und Google Flow werden also getrennt von älteren oder anderen Modellversionen gezählt. Das Ranking wird täglich neu berechnet.
- 1Fable 5unknown · 46× · 16. Aug. 2026100
- 2DeepSeek R1deepseek · 106× · 17. Aug. 202668
- 3SubQ 1.1 Smallsubquadratic · 16× · 17. Aug. 202667
- 4Daybreakopenai · 49× · 14. Aug. 202666
- 5MAI-Thinking-1microsoft · 55× · 14. Aug. 202665
- 6DeepSeek V3deepseek · 55× · 17. Aug. 202662
- 7o1openai · 52× · 13. Aug. 202658
- 8Google Flowgoogle · 24× · 17. Aug. 202655
- 9Fable 5fable · 7× · 14. Aug. 202649
- 10Inklingunknown · 13× · 01. Aug. 202647
- 11o3openai · 41× · 10. Aug. 202638
- 12DeepSeek V4 Pro GAdeepseek · 2× · 13. Aug. 202630
- 13BDH-CQpathway · 2× · 12. Aug. 202629
- 14o3 miniopenai · 2× · 12. Aug. 202628
- 15GPT-5.4 Thinkingopenai · 10× · 08. Aug. 202626
- 16Maple-Previewdeepgrove-ai · 2× · 05. Aug. 202623
- 17Maple-Previewdeepgrove · 2× · 05. Aug. 202623
- 18QwQalibaba · 9× · 17. Aug. 202622
- 19Tinkerthinking-machines-lab · 20× · 18. Juli 202621
- 20UniVR-34Bbytedance · 4× · 15. Juli 202618
- 21GPT-Realtime-2.1-miniopenai · 5× · 09. Juli 202617
- 22Laguna XS 2.1poolside · 3× · 24. Juli 202616
- 23Axiom Mathaxiom-math · 4× · 24. Juli 202615
- 24DeepSeek R1 671Bdeepseek · 2× · 23. Juli 202614
- 25Antidoomliquid-ai · 3× · 09. Juli 202611
- 26TabPFNprior-labs · 3× · 27. Juli 202610
- 27Reflectionreflection · 5× · 25. Juli 202610
- 28Sol Proopenai · 3× · 04. Juli 20269
- 29AlphaProof Nexusgoogle-deepmind · 17× · 13. Juni 20269
- 30Gemini 3 Deep Thinkgoogle · 2× · 18. Juli 20266
- 31GPT-5.5 Thinkingopenai · 13× · 07. Juni 20266
- 32TML Interactionthinking-machines · 19× · 23. Mai 20265
- 33Harness-1university-of-illinois-at-urbana-champaign-uc-berkeley-chroma · 4× · 13. Juni 20264
- 34Interaction Modelthinking-machines-lab · 13× · 14. Mai 20263
- 35Vibethinkerweibo · 2× · 18. Juni 20263
- 36Deepseek R1 Distill Qwendeepseek · 3× · 16. Juni 20263
- 37OpenAI reasoning modelopenai · 3× · 18. Juni 20262
- 38MiroThinker-1.7miromind · 8× · 14. Mai 20262
- 39Step-Audio-R1.5step-fun · 4× · 27. Mai 20262
- 40Kimi K2 Thinkingmoonshot-ai · 8× · 17. Juni 20262
- 41Perceptron Mk1perceptron · 5× · 14. Mai 20261
- 42SID-1unknown · 3× · 22. Mai 20261
- 43Interaction Modelsthinking-machines · 4× · 20. Mai 20261
- 44AlphaProofgoogle-deepmind · 4× · 27. Mai 20261
- 45Reasonixdeepseek · 2× · 27. Mai 20261
- 46Step-Audio-R1step-fun · 2× · 27. Mai 20261
- 47Step-Audio-R1.1step-fun · 2× · 27. Mai 20261
- 48Starship V3spacex · 2× · 27. Mai 20261
- 49Ring-2.6-1Tunknown · 2× · 20. Mai 20261
- 50Ponderponder · 2× · 16. Mai 20261