AGENTKERNEL
- Home
- Leaderboard
- Agents
Agents: the best AI models in October 2026
As of Oct 09, 2026, Claude Fable 5.1 leads the “Agents” ranking, ahead of Claude Opus 5 and Qwen3.8 Max (0902).
Best model with open weights: GLM 5.3 (53.1).
Cheapest strong option: GLM 5.3 with 53.1 points, at most five points behind the leader, for $1.79 a month at 1M input and 250,000 output tokens.
Open in the interactive comparisonAgents ranking
Index points, not a percentage. A gap of one or two points is hardly noticeable in everyday use.
| Rank | Model | Licence | Points | Input / output per 1M |
|---|---|---|---|---|
| 1 | Claude Fable 5.1Anthropic | Proprietary | 57.9 | $10.00 / $50.00 |
| 2 | Claude Opus 5Anthropic | Proprietary | 56.5 | $5.00 / $25.00 |
| 3 | Qwen3.8 Max (0902)Qwen | Proprietary | 56 | $2.00 / $6.00 |
| 4 | Muse Spark 1.3Meta | Proprietary | 55.5 | $1.25 / $4.25 |
| 5 | GLM 5.3Z.ai | Open | 53.1 | $0.04 / $7.00 |
| 6 | Grok 4.6xAI | Proprietary | 53 | $2.00 / $6.00 |
| 7 | GPT-6 AstraOpenAI | Proprietary | 51 | $10.00 / $50.00 |
| 8 | GLM 5.3 FlashZ.ai | Open | 50.9 | $0.15 / $0.50 |
| 9 | Claude Fable 5Anthropic | Proprietary | 50.7 | $10.00 / $50.00 |
| 10 | GPT-5.6 SolOpenAI | Proprietary | 50.2 | $2.00 / $10.00 |
| 11 | Qwen3.8 2.4T A95BQwen | Open | 50.1 | $2.00 / $6.00 |
| 12 | Kimi K3Moonshot AI | Open | 50 | $0.82 / $13.50 |
| 13 | DeepSeek V4 Flash Vision ExpDeepSeek | Open | 47.5 | $0.216 / $0.647 |
| 14 | Qwen3.8 27BQwen | Open | 45.8 | $0.425 / $2.55 |
| 15 | Claude Sonnet 5Anthropic | Proprietary | 43.6 | $2.00 / $10.00 |