Value ranking
Best Value LLMs for Agentic & Tool Use in October 2026
As of October 2026, GPT-6 Luna gives the most agentic & tool use performance per dollar: 33.3 points at a blended $0.20 per million tokens. GLM-5.3-Flash is next.
Last verified
Category score per dollar of blended API price, among models in the top half of the agentic ranking.
- GPT-6 Luna 166.6 pts/$
- GLM-5.3-Flash 144.2 pts/$
- Grok 4.1 Fast 132.1 pts/$
- DeepSeek V4.1 Flash 118.8 pts/$
- DeepSeek-V3.2-Exp 112.8 pts/$
- Qwen3 8B 97.4 pts/$
- GPT-5.6 Luna 76.5 pts/$
- MiMo-V2.6-Pro 69.0 pts/$
- MiniMax-M2.5 57.8 pts/$
- GPT-4.1 mini 47.5 pts/$
Full ranking
| # | Model | Score | Input $/M | Output $/M | Points per $ |
|---|---|---|---|---|---|
| 1 | GPT-6 Luna | 33.3 | $0.10 | $0.50 | 166.6 |
| 2 | GLM-5.3-Flash | 34.2 | $0.15 | $0.50 | 144.2 |
| 3 | Grok 4.1 Fast | 36.3 | $0.20 | $0.50 | 132.1 |
| 4 | DeepSeek V4.1 Flash | 31.2 | $0.15 | $0.60 | 118.8 |
| 5 | DeepSeek-V3.2-Exp | 32.7 | $0.26 | $0.38 | 112.8 |
| 6 | Qwen3 8B | 30.2 | $0.18 | $0.70 | 97.4 |
| 7 | GPT-5.6 Luna | 34.4 | $0.20 | $1.20 | 76.5 |
| 8 | MiMo-V2.6-Pro | 37.5 | $0.43 | $0.87 | 69.0 |
| 9 | MiniMax-M2.5 | 30.4 | $0.30 | $1.20 | 57.8 |
| 10 | GPT-4.1 mini | 33.3 | $0.40 | $1.60 | 47.5 |
| 11 | GPT-5 Mini | 31.1 | $0.25 | $2 | 45.3 |
| 12 | Kimi K2.5 | 34.2 | $0.45 | $2.25 | 38.0 |
| 13 | Gemini 2.5 Flash | 30.8 | $0.30 | $2.50 | 36.2 |
| 14 | Gemini 3 Flash Preview | 38.7 | $0.50 | $3 | 34.4 |
| 15 | DeepSeek-R1 | 30.7 | $0.50 | $2.15 | 33.6 |
| 16 | DeepSeek V4 Pro | 32.8 | $0.66 | $1.98 | 33.1 |
| 17 | GLM-4.6 | 32.3 | $0.60 | $2.20 | 32.3 |
| 18 | Kimi K2 (Jul 2025) | 32.4 | $0.57 | $2.30 | 32.3 |
| 19 | Qwen3.8 27B | 32.9 | $0.99 | $1.49 | 29.5 |
| 20 | Gemini 3.7 Flash | 42.1 | $0.75 | $3.75 | 28.1 |
| 21 | Gemini 3.8 Flash | 41.8 | $0.75 | $3.75 | 27.9 |
| 22 | Qwen3 235B-A22B | 33.9 | $0.70 | $2.80 | 27.7 |
| 23 | Qwen3 32B | 32.6 | $0.70 | $2.80 | 26.6 |
| 24 | Qwen3.5 397B-A17B | 33.3 | $0.60 | $3.60 | 24.7 |
| 25 | Grok 4.20 (Non-Reasoning) | 34.4 | $1.25 | $2.50 | 22.0 |
| 26 | Gemini 3.6 Flash | 32.3 | $0.75 | $3.75 | 21.6 |
| 27 | GLM-5 | 31.1 | $1 | $3.20 | 20.0 |
| 28 | Muse Spark 1.3 | 38.6 | $1.25 | $4.25 | 19.3 |
| 29 | o4-mini | 32.6 | $1.10 | $4.40 | 16.9 |
| 30 | GLM-5.3 | 36.4 | $1.40 | $4.40 | 16.9 |
| 31 | Claude Haiku 4.5 | 33.6 | $1 | $5 | 16.8 |
| 32 | Muse Spark 1.1 | 30.8 | $1.25 | $4.25 | 15.4 |
| 33 | Qwen3.8 Max | 45.4 | $2 | $6 | 15.1 |
| 34 | GLM-5.2 | 32.4 | $1.40 | $4.40 | 15.1 |
| 35 | Grok 4.5 | 44.4 | $2 | $6 | 14.8 |
| 36 | Grok 4.6 | 39.4 | $2 | $6 | 13.1 |
| 37 | Grok 4.7 | 36.7 | $2 | $6 | 12.2 |
| 38 | Claude Sonnet 5.5 | 45.0 | $2 | $10 | 11.3 |
| 39 | GPT-5.1-Codex | 38.0 | $1.25 | $10 | 11.1 |
| 40 | Claude Sonnet 5 | 42.8 | $2 | $10 | 10.7 |
| 41 | GPT-5.3 Codex | 48.0 | $1.75 | $14 | 10.0 |
| 42 | GPT-4.1 | 34.7 | $2 | $8 | 9.9 |
| 43 | GPT-6.1 Sol | 39.6 | $2 | $10 | 9.9 |
| 44 | o3 | 34.5 | $2 | $8 | 9.9 |
| 45 | GPT-5 | 33.1 | $1.25 | $10 | 9.6 |
| 46 | GPT-5.1 | 32.7 | $1.25 | $10 | 9.5 |
| 47 | GPT-6 Sol | 37.2 | $2 | $10 | 9.3 |
| 48 | GPT-5-Codex | 31.0 | $1.25 | $10 | 9.0 |
| 49 | GPT-5.6 Terra | 40.1 | $2 | $12 | 8.9 |
| 50 | GPT-5.2 Codex | 41.0 | $1.75 | $14 | 8.5 |
| 51 | Gemini 3.1 Pro Preview | 37.7 | $2 | $12 | 8.4 |
| 52 | GPT-5.2 | 40.2 | $1.75 | $14 | 8.3 |
| 53 | GPT-5.4 | 46.5 | $2.50 | $15 | 8.3 |
| 54 | Command A | 35.9 | $2.50 | $10 | 8.2 |
| 55 | Kimi K3 | 41.8 | $3 | $15 | 7.0 |
| 56 | Claude Sonnet 4.6 | 39.1 | $3 | $15 | 6.5 |
| 57 | Claude Sonnet 4 | 38.5 | $3 | $15 | 6.4 |
| 58 | Claude Sonnet 4.5 | 38.3 | $3 | $15 | 6.4 |
| 59 | GPT-5.6 Sol | 50.3 | $4 | $20 | 6.3 |
| 60 | Claude Opus 5.5 | 45.3 | $4 | $20 | 5.7 |
| 61 | Claude Opus 5 | 55.6 | $5 | $25 | 5.6 |
| 62 | Claude Opus 4.6 | 51.1 | $5 | $25 | 5.1 |
| 63 | Claude Opus 4.7 | 47.9 | $5 | $25 | 4.8 |
| 64 | Claude Opus 4.8 | 47.6 | $5 | $25 | 4.8 |
| 65 | Claude Opus 4.5 | 47.3 | $5 | $25 | 4.7 |
| 66 | GPT-5.5 | 50.7 | $5 | $30 | 4.5 |
| 67 | Claude Fable 5 | 54.0 | $10 | $50 | 2.7 |
| 68 | GPT-6 Astra | 52.9 | $10 | $50 | 2.6 |
| 69 | Claude Fable 5.1 | 50.7 | $10 | $50 | 2.5 |
| 70 | Claude Opus 4.1 | 35.0 | $15 | $75 | 1.2 |
| 71 | Claude Opus 4 | 34.8 | $15 | $75 | 1.2 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Compare the leaders
Frequently asked questions
What is the most cost-effective AI model for agents?
As of October 2026, GPT-6 Luna gives the most agentic & tool use performance per dollar: 33.3 points at a blended $0.20 per million tokens. GLM-5.3-Flash is next.
What are the top 5 in this ranking?
GPT-6 Luna, GLM-5.3-Flash, Grok 4.1 Fast, DeepSeek V4.1 Flash, DeepSeek-V3.2-Exp, in that order, as of October 2026.
How is this list ranked?
By score per dollar: the model’s score divided by its blended API price (three parts input to one part output), among models in the top half by score.