Value ranking
Best Value LLMs for Coding in October 2026
As of October 2026, Mercury 2.5 gives the most coding performance per dollar: 39.5 points at a blended $0.07 per million tokens. Nemotron 3 Nano 30B A3B is next.
Last verified
Category score per dollar of blended API price, among models in the top half of the coding ranking.
- Mercury 2.5 585.2 pts/$
- Nemotron 3 Nano 30B A3B 462.9 pts/$
- Nemotron 3.5 Lightning 461.6 pts/$
- Granite 4.2 8B 377.1 pts/$
- Gemma 4 26B A4B IT 365.4 pts/$
- Hy3 324.4 pts/$
- MiMo-V2.6-Flash 305.3 pts/$
- Step 3.5 Flash 282.4 pts/$
- GLM-4.7-Flash 280.2 pts/$
- GPT-6 Luna 277.4 pts/$
Full ranking
| # | Model | Score | Input $/M | Output $/M | Points per $ |
|---|---|---|---|---|---|
| 1 | Mercury 2.5 | 39.5 | $0.04 | $0.15 | 585.2 |
| 2 | Nemotron 3 Nano 30B A3B | 40.5 | $0.05 | $0.20 | 462.9 |
| 3 | Nemotron 3.5 Lightning | 40.4 | $0.05 | $0.20 | 461.6 |
| 4 | Granite 4.2 8B | 40.5 | $0.06 | $0.25 | 377.1 |
| 5 | Gemma 4 26B A4B IT | 39.0 | $0.0675 | $0.23 | 365.4 |
| 6 | Hy3 | 46.8 | $0.0825 | $0.33 | 324.4 |
| 7 | MiMo-V2.6-Flash | 53.4 | $0.14 | $0.28 | 305.3 |
| 8 | Step 3.5 Flash | 42.4 | $0.10 | $0.30 | 282.4 |
| 9 | GLM-4.7-Flash | 40.6 | $0.06 | $0.40 | 280.2 |
| 10 | GPT-6 Luna | 55.5 | $0.10 | $0.50 | 277.4 |
| 11 | Gemma 4 31B IT | 42.3 | $0.09 | $0.34 | 277.2 |
| 12 | Devstral Small 2505 | 38.9 | $0.10 | $0.30 | 259.5 |
| 13 | MiMo-V2.5 | 43.9 | $0.14 | $0.28 | 251.1 |
| 14 | MiMo-V2-Omni | 43.3 | $0.14 | $0.28 | 247.7 |
| 15 | Claude Haiku 5.5 | 49.2 | $0.10 | $0.50 | 245.9 |
| 16 | Nemotron 3 Super | 39.4 | $0.08 | $0.45 | 228.2 |
| 17 | GLM-5.3-Flash | 53.1 | $0.15 | $0.50 | 223.5 |
| 18 | DeepSeek V4.1 Flash | 52.9 | $0.15 | $0.60 | 201.6 |
| 19 | DeepSeek V4 Flash | 47.9 | $0.15 | $0.60 | 182.6 |
| 20 | DeepSeek-V3.2-Exp | 46.5 | $0.26 | $0.38 | 160.4 |
| 21 | GPT-5.6 Luna | 54.5 | $0.20 | $1.20 | 121.1 |
| 22 | DeepSeek-V3 | 42.3 | $0.24 | $0.90 | 104.5 |
| 23 | MiMo-V2.6-Pro | 55.5 | $0.43 | $0.87 | 102.2 |
| 24 | Nvidia Llama 3.3 Nemotron Super 49b v1.5 | 39.8 | $0.40 | $0.40 | 99.4 |
| 25 | Step 3.7 Flash | 40.0 | $0.18 | $1.11 | 96.0 |
| 26 | DeepSeek-V3.1 | 40.3 | $0.25 | $0.95 | 94.9 |
| 27 | GPT-5.4 nano | 43.6 | $0.20 | $1.25 | 94.3 |
| 28 | DeepSeek-V3.1-Terminus | 42.0 | $0.27 | $1 | 92.9 |
| 29 | MiniMax-M2.5 | 48.1 | $0.30 | $1.20 | 91.5 |
| 30 | GLM-4.6V | 40.9 | $0.30 | $0.90 | 90.8 |
| 31 | MiMo-V2.5-Pro | 47.4 | $0.43 | $0.87 | 87.2 |
| 32 | MiMo-V2-Pro | 43.8 | $0.43 | $0.87 | 80.5 |
| 33 | MiniMax-M3 | 41.8 | $0.30 | $1.20 | 79.6 |
| 34 | MiniMax-M2.7 | 41.8 | $0.30 | $1.20 | 79.5 |
| 35 | MiniMax-M2.1 | 40.4 | $0.30 | $1.20 | 76.9 |
| 36 | Solar Pro4 | 40.1 | $0.30 | $1.20 | 76.4 |
| 37 | MiniMax-M2 | 39.3 | $0.30 | $1.20 | 74.8 |
| 38 | Inkling-Small | 43.6 | $0.45 | $1.20 | 68.4 |
| 39 | Qwen Plus | 38.9 | $0.40 | $1.20 | 64.8 |
| 40 | GPT-5 Mini | 40.1 | $0.25 | $2 | 58.4 |
| 41 | Kimi K2.5 | 48.8 | $0.45 | $2.25 | 54.2 |
| 42 | DeepSeek V4 Pro | 52.4 | $0.66 | $1.98 | 53.0 |
| 43 | DeepSeek-R1 | 46.3 | $0.50 | $2.15 | 50.7 |
| 44 | Gemini 3.5 Flash Lite | 42.4 | $0.30 | $2.50 | 49.8 |
| 45 | Qwen3-Next 80B-A3B Instruct | 42.5 | $0.50 | $2 | 48.6 |
| 46 | Nova 2 Lite | 40.7 | $0.30 | $2.50 | 47.9 |
| 47 | DeepSeek-V3.2-Speciale | 40.4 | $0.58 | $1.68 | 47.2 |
| 48 | Qwen3.5 27B | 38.9 | $0.30 | $2.40 | 47.1 |
| 49 | Mistral Large 4 | 48.6 | $0.68 | $2.09 | 47.1 |
| 50 | Hy4 preview | 51.6 | $0.75 | $2.25 | 45.8 |
| 51 | Qwen3.8 27B | 50.5 | $0.99 | $1.49 | 45.3 |
| 52 | Gemini 3 Flash Preview | 50.9 | $0.50 | $3 | 45.3 |
| 53 | GLM-4.7 | 44.0 | $0.60 | $2.20 | 44.0 |
| 54 | GLM-4.5V | 39.5 | $0.60 | $1.80 | 43.9 |
| 55 | Kimi K2 (Jul 2025) | 42.4 | $0.57 | $2.30 | 42.3 |
| 56 | MiniMax M1 | 39.9 | $0.55 | $2.20 | 41.4 |
| 57 | GLM-4.5 | 41.4 | $0.60 | $2.20 | 41.4 |
| 58 | GLM-4.6 | 40.1 | $0.60 | $2.20 | 40.1 |
| 59 | Gemini 3.8 Flash | 59.2 | $0.75 | $3.75 | 39.5 |
| 60 | Seed 2.0 Pro | 43.5 | $0.50 | $3 | 38.7 |
| 61 | Gemini 3.7 Flash | 56.2 | $0.75 | $3.75 | 37.5 |
| 62 | Step 5 Preview | 51.8 | $1 | $2.70 | 36.3 |
| 63 | Qwen3.6 Plus | 40.8 | $0.50 | $3 | 36.3 |
| 64 | Qwen3 235B-A22B | 44.3 | $0.70 | $2.80 | 36.1 |
| 65 | Qwen3.5 122B-A10B | 39.1 | $0.40 | $3.20 | 35.6 |
| 66 | Qwen3-VL 235B-A22B | 42.4 | $0.70 | $2.80 | 34.6 |
| 67 | Grok Build 0.1 | 43.1 | $1 | $2 | 34.5 |
| 68 | Gemini 3.6 Flash | 50.0 | $0.75 | $3.75 | 33.4 |
| 69 | GLM-5 | 49.0 | $1 | $3.20 | 31.6 |
| 70 | Qwen3.5 397B-A17B | 42.0 | $0.60 | $3.60 | 31.1 |
| 71 | Kimi K2.6 | 50.7 | $0.95 | $4 | 29.6 |
| 72 | Qwen3.6 27B | 39.1 | $0.60 | $3.60 | 29.0 |
| 73 | Muse Spark 1.3 | 56.6 | $1.25 | $4.25 | 28.3 |
| 74 | GLM-5.3 | 59.5 | $1.40 | $4.40 | 27.7 |
| 75 | Grok 4.20 Multi-Agent | 43.0 | $1.25 | $2.50 | 27.6 |
| 76 | Grok 4.20 (Non-Reasoning) | 42.1 | $1.25 | $2.50 | 26.9 |
| 77 | GPT-5.4 mini | 45.2 | $0.75 | $4.50 | 26.8 |
| 78 | Grok 4.3 | 41.6 | $1.25 | $2.50 | 26.6 |
| 79 | Muse Spark 1.1 | 51.3 | $1.25 | $4.25 | 25.6 |
| 80 | Kimi K2.7 Code | 42.9 | $0.95 | $4 | 25.0 |
| 81 | Muse Spark 1.2 | 49.2 | $1.25 | $4.25 | 24.6 |
| 82 | GLM-5.2 | 51.3 | $1.40 | $4.40 | 23.8 |
| 83 | GLM-5.1 | 48.7 | $1.40 | $4.40 | 22.6 |
| 84 | GLM-5V-Turbo | 42.1 | $1.20 | $4 | 22.1 |
| 85 | Claude Haiku 4.5 | 44.0 | $1 | $5 | 22.0 |
| 86 | o4-mini | 40.9 | $1.10 | $4.40 | 21.3 |
| 87 | o3-mini | 40.8 | $1.10 | $4.40 | 21.2 |
| 88 | Grok 4.6 | 58.5 | $2 | $6 | 19.5 |
| 89 | Grok 4.7 | 58.0 | $2 | $6 | 19.3 |
| 90 | Qwen3 Max | 43.0 | $1.20 | $6 | 17.9 |
| 91 | Qwen3.8 Max | 53.5 | $2 | $6 | 17.8 |
| 92 | Grok 4.5 | 52.2 | $2 | $6 | 17.4 |
| 93 | Claude Sonnet 5.5 | 67.3 | $2 | $10 | 16.8 |
| 94 | Qwen3.6 Max Preview | 48.7 | $1.30 | $7.80 | 16.7 |
| 95 | GPT-6.1 Sol | 63.2 | $2 | $10 | 15.8 |
| 96 | GPT-6 Sol | 60.1 | $2 | $10 | 15.0 |
| 97 | Gemini 3.5 Flash | 49.4 | $1.50 | $9 | 14.6 |
| 98 | GPT-5 | 50.3 | $1.25 | $10 | 14.6 |
| 99 | Magistral Medium | 39.1 | $2 | $5 | 14.2 |
| 100 | Claude Sonnet 5 | 55.5 | $2 | $10 | 13.9 |
| 101 | GPT-5.1 | 46.4 | $1.25 | $10 | 13.5 |
| 102 | Qwen3.7 Max | 50.4 | $2.50 | $7.50 | 13.4 |
| 103 | o3 | 46.8 | $2 | $8 | 13.4 |
| 104 | GPT-5.6 Terra | 57.7 | $2 | $12 | 12.8 |
| 105 | Gemini 2.5 Pro | 42.4 | $1.25 | $10 | 12.3 |
| 106 | GPT-5-Codex | 42.4 | $1.25 | $10 | 12.3 |
| 107 | GPT-5.1-Codex | 41.9 | $1.25 | $10 | 12.2 |
| 108 | GPT-5.2 | 51.6 | $1.75 | $14 | 10.7 |
| 109 | Kimi K3 | 61.0 | $3 | $15 | 10.2 |
| 110 | GPT-5.3 Codex | 48.6 | $1.75 | $14 | 10.1 |
| 111 | GPT-5.2 Codex | 45.5 | $1.75 | $14 | 9.5 |
| 112 | Gemini 3.1 Pro Preview | 42.5 | $2 | $12 | 9.4 |
| 113 | GPT-5.4 | 52.6 | $2.50 | $15 | 9.4 |
| 114 | Claude Opus 5.5 | 71.9 | $4 | $20 | 9.0 |
| 115 | GPT-5.3 Chat | 41.4 | $1.75 | $14 | 8.6 |
| 116 | GPT-5.6 Sol | 65.1 | $4 | $20 | 8.1 |
| 117 | Claude Sonnet 4.5 | 47.3 | $3 | $15 | 7.9 |
| 118 | Claude Sonnet 4.6 | 46.3 | $3 | $15 | 7.7 |
| 119 | Claude Sonnet 4 | 43.5 | $3 | $15 | 7.2 |
| 120 | Claude Opus 5 | 67.5 | $5 | $25 | 6.7 |
| 121 | Claude Opus 4.8 | 59.9 | $5 | $25 | 6.0 |
| 122 | Claude Opus 4.7 | 59.6 | $5 | $25 | 6.0 |
| 123 | Claude Opus 4.6 | 57.2 | $5 | $25 | 5.7 |
| 124 | Claude Opus 4.5 | 54.8 | $5 | $25 | 5.5 |
| 125 | GPT-5.5 | 58.2 | $5 | $30 | 5.2 |
| 126 | Claude Fable 5.1 | 74.7 | $10 | $50 | 3.7 |
| 127 | GPT-6 Astra | 73.7 | $10 | $50 | 3.7 |
| 128 | Claude Fable 5 | 70.6 | $10 | $50 | 3.5 |
| 129 | o1 | 46.1 | $15 | $60 | 1.8 |
| 130 | o3-pro | 55.5 | $20 | $80 | 1.6 |
| 131 | Claude Opus 4 | 47.2 | $15 | $75 | 1.6 |
| 132 | Claude Opus 4.1 | 44.4 | $15 | $75 | 1.5 |
| 133 | GPT-5 Pro | 44.0 | $15 | $120 | 1.1 |
| 134 | GPT-5.4 Pro | 43.2 | $30 | $180 | 0.6 |
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 coding?
As of October 2026, Mercury 2.5 gives the most coding performance per dollar: 39.5 points at a blended $0.07 per million tokens. Nemotron 3 Nano 30B A3B is next.
What are the top 5 in this ranking?
Mercury 2.5, Nemotron 3 Nano 30B A3B, Nemotron 3.5 Lightning, Granite 4.2 8B, Gemma 4 26B A4B IT, 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.