Best under $0.50/M
Highest quality for budget workloads, below fifty cents per million tokens.
101 models qualify · data as of 9 Sept 2026
- 1GLM-5.3-Flash61
- 2GPT-5.6 Luna59.3
- 3DeepSeek V4 Flash59.3
- 4Qwen3.8-Flash-Next59.3
- 5Qwen3.5 397B-A17B58.6
- 6MiMo-V2.558.3
- 7Deepseek v4 Flash Vision57.4
- 8DeepSeek V3.255.9
- 9Gemma-4-31B-IT55.4
- 10Step 3.5 Flash54.9
- 11Step 3.7 Flash54.6
- 12Gemma 4 26B A4B54
- 13Ling 3.0 Flash54
- 14GPT-5.4 nano53.9
- 15Nemotron 3 Super53.6
101 of 101
| # | |||||
|---|---|---|---|---|---|
| 1 |  GLM-5.3-FlashZhipu AIopen ↗ | 61 | $0.119/M | $0.119 | 72 |
| 2 |  GPT-5.6 LunaxhighOpenAI | 59.3 | $0.450/M | $0.450 | 106 |
| 3 |  DeepSeek V4 FlashhighDeepSeekopen ↗ | 59.3 | $0.168/M | $0.168 | 118 |
| 4 |  Qwen3.8-Flash-NextAlibabaopen ↗ | 59.3 | $0.230/M | $0.230 | 64 |
| 5 | Qwen3.5 397B-A17BAlibabaopen ↗ | 58.6 | $0.387/M | $0.387 | 81 |
| 6 | MiMo-V2.5Xiaomiopen ↗ | 58.3 | $0.175/M | $0.175 | 54 |
| 7 | Deepseek v4 Flash VisionDeepSeek | 57.4 | $0.175/M | $0.175 | 119 |
| 8 | DeepSeek V3.2thinkingDeepSeekopen ↗ | 55.9 | $0.315/M | $0.315 | 30 |
| 9 |  Gemma-4-31B-ITNVIDIAopen ↗ | 55.4 | $0.205/M | $0.205 | 35 |
| 10 |  Step 3.5 FlashStepFunopen ↗ | 54.9 | $0.150/M | $0.150 | 216 |
| 11 | Step 3.7 FlashStepFunopen ↗ | 54.6 | $0.416/M | $0.416 | 86 |
| 12 |  Gemma 4 26B A4BGoogleopen ↗ | 54 | $0.190/M | $0.190 | 30 |
| 13 | Ling 3.0 FlashInclusionAIopen ↗ | 54 | $0.115/M | $0.115 | 290 |
| 14 | GPT-5.4 nanoOpenAI | 53.9 | $0.463/M | $0.463 | 169 |
| 15 | Nemotron 3 SuperNVIDIAopen ↗ | 53.6 | $0.350/M | $0.350 | 96 |
| 16 | MiMo-V2-FlashthinkingXiaomiopen ↗ | 53.4 | $0.150/M | $0.150 | – |
| 17 | Gemma 4 12BGoogleopen ↗ | 52.1 | $0.150/M | $0.150 | 107 |
| 18 |  Granite 4.2 30BIBMopen ↗ | 51.3 | $0.283/M | $0.283 | 76 |
| 19 | GPT-5 nanoOpenAI | 51.1 | $0.138/M | $0.138 | 176 |
| 20 | Qwen3-Next 80B A3BthinkingAlibabaopen ↗ | 51 | $0.466/M | $0.466 | 190 |
| 21 | GLM-4.7-FlashZhipu AIopen ↗ | 50 | $0.153/M | $0.153 | 97 |
| 22 |  Mercury 2Inception | 49.5 | $0.375/M | $0.375 | 744 |
| 23 | HyperNova 60B 2605 (high, based on gpt-oss-120b)Multiverse Computingopen | 49.4 | $0.065/M | $0.065 | 345 |
| 24 | DeepSeek R1DeepSeekopen ↗ | 48.9 | $0.315/M | $0.315 | 19 |
| 25 | Gemini 2.5 Flash Lite 09thinkingGoogle | 48.8 | $0.175/M | $0.175 | – |
| 26 | Trinity LargethinkingArcee AI | 48.7 | $0.413/M | $0.413 | 321 |
| 27 | Granite 4.2 8BIBMopen ↗ | 48.6 | $0.108/M | $0.108 | 120 |
| 28 | nemotron-3-nano-30b-a3bthinkingNVIDIAopen ↗ | 48.6 | $0.088/M | $0.088 | 231 |
| 29 | Qwen PlusAlibaba | 48.4 | $0.158/M | $0.158 | 37 |
| 30 | ByteDance-Seed/Seed-OSS-36B-InstructNVIDIA | 48 | $0.300/M | $0.300 | 26 |
| 31 | Qwen3.5-9Bno reasoningAlibabaopen ↗ | 47.7 | $0.190/M | $0.190 | 89 |
| 32 | Granite 4.2 3BIBMopen ↗ | 47.6 | $0.053/M | $0.053 | 215 |
| 33 | gpt-oss-20bOpenAIopen ↗ | 47.1 | $0.085/M | $0.085 | 192 |
| 34 | Qwen3.5-4BAlibaba | 47.1 | $0.060/M | $0.060 | 27 |
| 35 |  Mistral Small 4Mistral AIopen | 47.1 | $0.262/M | $0.262 | 168 |
| 36 | Gemini 2.0 Flash-LiteGoogle | 47 | $0.131/M | $0.131 | – |
| 37 | Gemini 2.5 Flash-LitethinkingGoogle | 46.7 | $0.175/M | $0.175 | 378 |
| 38 | GLM-4.5-AirZhipu AIopen ↗ | 46 | $0.425/M | $0.425 | 54 |
| 39 | Nemotron 3 Nano Omni 30B A3bNVIDIAopen | 46 | $0.158/M | $0.158 | – |
| 40 | QwQ-32BAlibabaopen ↗ | 45.7 | $0.431/M | $0.431 | – |
| 41 | Qwen3.5 Omni FlashAlibaba | 45.6 | $0.275/M | $0.275 | 235 |
| 42 | Qwen3 30B A3BAlibabaopen ↗ | 45.4 | $0.215/M | $0.215 | 108 |
| 43 | Gemma 4 E4BGoogleopen ↗ | 45.4 | $0.040/M | $0.040 | 41 |
| 44 | Ring Flash 2.0Ant-group | 45.2 | $0.248/M | $0.248 | – |
| 45 | GLM-4.6VthinkingZhipu AIopen ↗ | 45.2 | $0.450/M | $0.450 | 56 |
| 46 | Gemini 2.0 FlashthinkingGoogle | 45.1 | $0.175/M | $0.175 | – |
| 47 | Ling Flash 2.0Ant-group | 45 | $0.248/M | $0.248 | 7 |
| 48 |  ERNIE 4.5Baiduopen | 44.8 | $0.485/M | $0.485 | – |
| 49 | Llama 3.3 Nemotron Super 49B v1.5NVIDIAopen | 44.7 | $0.400/M | $0.400 | – |
| 50 |  Solar Pro 3Upstage | 44.6 | $0.262/M | $0.262 | 147 |
| 51 | Qwen3.6 FlashAlibaba | 44.6 | $0.422/M | $0.422 | 100 |
| 52 | Qwen3-Coder 30B-A3B InstructAlibabaopen ↗ | 44.4 | $0.377/M | $0.377 | 86 |
| 53 | qwen3-coder-nextAlibabaopen ↗ | 44.1 | $0.450/M | $0.450 | 113 |
| 54 | DeepSeek V3DeepSeekopen ↗ | 43.8 | $0.478/M | $0.478 | – |
| 55 | Reka FlashReka | 43.1 | $0.350/M | $0.350 | – |
| 56 | DeepSeek-R1-Distill-Qwen-32BDeepSeekopen ↗ | 42.8 | $0.440/M | $0.440 | – |
| 57 |  Llama 4 MaverickMetaopen ↗ | 42.6 | $0.400/M | $0.400 | 91 |
| 58 | Llama Nemotron Super 49B v1.5NVIDIAopen | 42.6 | $0.400/M | $0.400 | 35 |
| 59 | Devstral SmallMistral AIopen | 42.5 | $0.150/M | $0.150 | – |
| 60 | Qwen2.5-Coder 32B InstructAlibabaopen ↗ | 42.2 | $0.431/M | $0.431 | 27 |