AI-modeller
Rangering fra Arena.ai, blindtester og agentoppgaver.
Agentiske oppgaver med flerstegsresonnering og verktøybruk.
| # | Modell | Leverandør | Netto forbedring |
|---|---|---|---|
| 01 | Claude Fable 5.1 (Max) | Anthropic | +13,85 % |
| 02 | GPT 6 Astra (Max) | OpenAI | +12,39 % |
| 03 | Claude Opus 5 (High) | Anthropic | +11,06 % |
| 04 | Claude Opus 5 (Max) | Anthropic | +10,80 % |
| 05 | Claude Fable 5 (High) | Anthropic | +9,09 % |
| 06 | Claude Opus 4.8 (High) | Anthropic | +7,57 % |
| 07 | GPT 5.6 Sol (xHigh) | OpenAI | +7,51 % |
| 08 | Kimi K3 (Max) | Moonshot | +6,46 % |
| 09 | Claude Sonnet 5 (High) | Anthropic | +5,88 % |
| 10 | Hy4 preview | Tencent | +5,23 % |
| 11 | GPT 5.5 (xHigh) | OpenAI | +5,22 % |
| 12 | GLM 5.2 (Max) | Z.ai | +4,56 % |
| 13 | Muse Spark 1.3 (Max) | Meta | +4,23 % |
| 14 | DeepSeek V4 Pro (High) (0813) | DeepSeek | +4,14 % |
| 15 | Qwen3.8 Max | Alibaba | +3,97 % |
| 16 | Gemini 3.8 Flash (High) | +3,84 % | |
| 17 | Grok 4.5 | SpaceXAI | +3,83 % |
| 18 | Grok 4.6 (xHigh) | SpaceXAI | +3,12 % |
| 19 | GPT 5.5 | OpenAI | +2,80 % |
| 20 | GLM 5.3 (Max) | Z.ai | +2,62 % |
| 21 | Deepseek V4 Flash (High) (20260731) | DeepSeek | +2,17 % |
| 22 | GLM 5.3 Flash | Z.ai | +1,89 % |
| 23 | GPT 5.4 (High) | OpenAI | +1,54 % |
| 24 | GPT 5.6 Terra (xHigh) | OpenAI | +1,46 % |
| 25 | GPT 5.6 Luna (xHigh) | OpenAI | +0,73 % |
Data fra arena.ai. Agent-boardet oppdateres daglig, sist 14. september 2026. Kode-boardet er snapshot 16. juni 2026.