usefulHQ Study · 2026
The same month of AI costs $0.196 on DeepSeek vs $40 on Claude Opus, a 204× spread
We priced a typical assistant workload, 5 million input and 1 million output tokens a month, across the flagship models using live OpenRouter rates. The cheapest capable option, DeepSeek V4 Flash 0423, runs about $0.196 a month; the priciest, Claude Opus 5.5, about $40. For most tasks a mid-tier model does the job at a fraction of the frontier price, so matching the model to the task, not defaulting to the biggest name, is where the savings are.
Cheapest flagship
$0.196/mo
DeepSeek
Priciest flagship
$40/mo
Claude Opus
Cost spread
204×
same workload
Monthly cost by model (5M in + 1M out)
Method: live OpenRouter per-token rates × 5M input + 1M output tokens/month; batch and prompt-caching discounts excluded. Cite this: "usefulHQ LLM Cost Study 2026", tools.usefulhq.com/llm-pricing/. Updates from live prices. More usefulHQ studies →
Embed or cite this stat
Paste this into your page (a free link back to the data):
million
million
| Your cost/mo ▲ | Model | Provider | Input /1M | Output /1M | Context | |
|---|---|---|---|---|---|---|
| Free | Ling 3.1 Flash🧠 | inclusionAI | – | – | 262K | OpenRouter → |
| Free | Apodex 1.1 Mini (free)🧠 | Apodex | – | – | 262K | OpenRouter → |
| Free | Space Bunny Alpha👁🧠 | Other | – | – | 1M | OpenRouter → |
| Free | Ling 3.0 Flash Sante (free)🧠 | inclusionAI | – | – | 262K | OpenRouter → |
| Free | Qwen3.8 27B (free)👁🧠 | Qwen | – | – | 262K | OpenRouter → |
| Free | Dots3-Note Preview (free)👁🧠 | Dots Studio | – | – | 512K | OpenRouter → |
| Free | LFM2.5-2.6B (free)🧠 | LiquidAI | – | – | 66K | OpenRouter → |
| Free | Nemotron 3.5 Lightning (free)🧠 | NVIDIA | – | – | 1M | OpenRouter → |
| Free | Inkling Small (free)👁🧠 | Thinking Machines | – | – | 1.048576M | OpenRouter → |
| Free | Laguna S 2.1 (free)🧠 | Poolside | – | – | 262K | OpenRouter → |
| Free | Inkling (free)👁🧠 | Thinking Machines | – | – | 1.048576M | OpenRouter → |
| Free | Laguna XS 2.1 (free)🧠 | Poolside | – | – | 262K | OpenRouter → |
| Free | North Mini Code (free)🧠 | Cohere | – | – | 256K | OpenRouter → |
| Free | Nemotron 3.5 Content Safety (free)👁🧠 | NVIDIA | – | – | 128K | OpenRouter → |
| Free | Nemotron 3 Ultra (free)🧠 | NVIDIA | – | – | 1M | OpenRouter → |
| Free | Nemotron 3 Nano Omni (free)👁🧠 | NVIDIA | – | – | 256K | OpenRouter → |
| Free | Gemma 4 26B A4B (free)👁🧠 | – | – | 262K | OpenRouter → | |
| Free | Gemma 4 31B (free)👁🧠 | – | – | 262K | OpenRouter → | |
| Free | Lyria 3 Pro Preview👁 | – | – | 1.048576M | OpenRouter → | |
| Free | Lyria 3 Clip Preview👁 | – | – | 1.048576M | OpenRouter → | |
| Free | Nemotron 3 Super (free)🧠 | NVIDIA | – | – | 262K | OpenRouter → |
| Free | Free Models Router👁🧠 | Other | – | – | 200K | OpenRouter → |
| $0.125CHEAPEST | Mistral Nemo | Mistral | $0.019 | $0.03 | 131K | OpenRouter → |
| $0.1666 | Ling 3.0 Flash VL👁🧠 | inclusionAI | $0.021 | $0.0616 | 262K | OpenRouter → |
| $0.168 | Ling 3.0 Flash🧠 | inclusionAI | $0.021 | $0.063 | 262K | OpenRouter → |
| $0.18 | gpt-oss-20b🧠 | OpenAI | $0.018 | $0.09 | 131K | OpenRouter → |
| $0.196 | DeepSeek V4 Flash 0423🧠 | DeepSeek | $0.028 | $0.056 | 1.048576M | OpenRouter → |
| $0.197 | Granite 4.0 Micro | IBM | $0.017 | $0.112 | 131K | OpenRouter → |
| $0.225 | Nex-N2.5-Mini👁🧠 | Nex AGI | $0.025 | $0.1 | 262K | OpenRouter → |
| $0.232 | gpt-oss-20b (batch)🧠 | OpenAI | $0.024 | $0.112 | 131K | OpenRouter → |
| $0.25 | Llama 3 8B Lunaris | Sao10K | $0.04 | $0.05 | 8K | OpenRouter → |
| $0.28 | Qwen3.7 Flash👁🧠 | Qwen | $0.03 | $0.13 | 1M | OpenRouter → |
| $0.284 | gpt-oss-120b (batch)🧠 | OpenAI | $0.0296 | $0.136 | 131K | OpenRouter → |
| $0.3 | Schematron V2 Turbo | Inference.net | $0.03 | $0.15 | 128K | OpenRouter → |
| $0.315 | Nova Micro 1.0 | Amazon | $0.035 | $0.14 | 128K | OpenRouter → |
| $0.325 | GPT-5 Nano (batch)👁🧠 | OpenAI | $0.025 | $0.2 | 400K | OpenRouter → |
| $0.33 | Mistral Small 3 | Mistral | $0.05 | $0.08 | 33K | OpenRouter → |
| $0.33 | Llama 3.1 8B Instruct | Meta | $0.05 | $0.08 | 131K | OpenRouter → |
| $0.3332 | Ling 3.0 Flash Fin🧠 | inclusionAI | $0.042 | $0.1232 | 262K | OpenRouter → |
| $0.336 | Llama 3.2 1B Instruct | Meta | $0.027 | $0.201 | 60K | OpenRouter → |
| $0.3375 | Command R7B (12-2024) | Cohere | $0.0375 | $0.15 | 128K | OpenRouter → |
| $0.35 | Mercury 2.5🧠 | Inception | $0.04 | $0.15 | 260K | OpenRouter → |
| $0.35 | Gemma 3 4B👁 | $0.05 | $0.1 | 131K | OpenRouter → | |
| $0.355 | gpt-oss-120b🧠 | OpenAI | $0.037 | $0.17 | 131K | OpenRouter → |
| $0.397 | Hy-MT2-1.8B | Tencent | $0.044 | $0.177 | 8K | OpenRouter → |
| $0.4 | Gemma 3 12B👁 | $0.05 | $0.15 | 131K | OpenRouter → | |
| $0.42 | Laguna XS 2.1🧠 | Poolside | $0.06 | $0.12 | 262K | OpenRouter → |
| $0.4338 | Qwen3 30B A3B Instruct 2507 | Qwen | $0.0482 | $0.1931 | 262K | OpenRouter → |
| $0.45 | Solar Mini 4🧠 | Upstage | $0.05 | $0.2 | 524K | OpenRouter → |
| $0.45 | Nemotron 3 Nano 30B A3B🧠 | NVIDIA | $0.05 | $0.2 | 262K | OpenRouter → |
| $0.45 | Gemini 2.5 Flash Lite (batch)👁🧠 | $0.05 | $0.2 | 1.048576M | OpenRouter → | |
| $0.45 | GPT-4.1 Nano (batch)👁 | OpenAI | $0.05 | $0.2 | 1.047576M | OpenRouter → |
| $0.45 | Ministral 3 8B 2512 (batch)👁 | Mistral | $0.075 | $0.075 | 262K | OpenRouter → |
| $0.4675 | Nemotron 3.5 Lightning🧠 | NVIDIA | $0.0595 | $0.17 | 262K | OpenRouter → |
| $0.48 | Schematron V2 Small | Inference.net | $0.05 | $0.23 | 128K | OpenRouter → |
| $0.49 | Phi 4 | Microsoft | $0.07 | $0.14 | 16K | OpenRouter → |
| $0.5 | GPT-6 Luna Pro (batch)👁🧠 | OpenAI | $0.05 | $0.25 | 1.05M | OpenRouter → |
| $0.5 | GPT-6 Luna (batch)👁🧠 | OpenAI | $0.05 | $0.25 | 1.05M | OpenRouter → |
| $0.5 | GLM 5.3 Flash (batch)👁🧠 | Z.ai | $0.06 | $0.2 | 1.048576M | OpenRouter → |
| $0.51 | MythoMax 13B | Other | $0.08 | $0.11 | 8K | OpenRouter → |
| $0.54 | Nova Lite 1.0👁 | Amazon | $0.06 | $0.24 | 300K | OpenRouter → |
| $0.55 | Granite 4.2 8B🧠 | IBM | $0.06 | $0.25 | 131K | OpenRouter → |
| $0.5625 | Gemma 4 26B A4B 👁🧠 | $0.0675 | $0.225 | 262K | OpenRouter → | |
| $0.58 | Llama 3.2 3B Instruct | Meta | $0.05 | $0.33 | 131K | OpenRouter → |
| $0.585 | Qwen3.5-Flash👁🧠 | Qwen | $0.065 | $0.26 | 1M | OpenRouter → |
| $0.6 | Reka Edge👁🧠 | Other | $0.1 | $0.1 | 16K | OpenRouter → |
| $0.6 | Ministral 3 3B 2512👁 | Mistral | $0.1 | $0.1 | 131K | OpenRouter → |
| $0.625 | Nex-N2.5-Pro👁🧠 | Nex AGI | $0.075 | $0.25 | 262K | OpenRouter → |
| $0.63 | Laguna S 2.1🧠 | Poolside | $0.09 | $0.18 | 1.048576M | OpenRouter → |
| $0.63 | Qwen3 Coder 30B A3B Instruct | Qwen | $0.07 | $0.28 | 262K | OpenRouter → |
| $0.65 | Qwen3.5-9B👁🧠 | Qwen | $0.1 | $0.15 | 262K | OpenRouter → |
| $0.65 | GPT-5 Nano👁🧠 | OpenAI | $0.05 | $0.4 | 400K | OpenRouter → |
| $0.665 | Hy-MT2-30B-A3B | Tencent | $0.074 | $0.295 | 8K | OpenRouter → |
| $0.665 | Hy-MT2-7B | Tencent | $0.074 | $0.295 | 8K | OpenRouter → |
| $0.675 | GLM Flash Latest👁🧠 | Z.ai | $0.035 | $0.5 | 1.048576M | OpenRouter → |
| $0.675 | Mistral Small 4 (batch)👁🧠 | Mistral | $0.075 | $0.3 | 262K | OpenRouter → |
| $0.675 | Seed 1.6 Flash👁🧠 | ByteDance Seed | $0.075 | $0.3 | 262K | OpenRouter → |
| $0.675 | gpt-oss-safeguard-20b🧠 | OpenAI | $0.075 | $0.3 | 131K | OpenRouter → |
| $0.675 | GPT-4o-mini (batch)👁 | OpenAI | $0.075 | $0.3 | 128K | OpenRouter → |
| $0.68 | Qwen3 32B🧠 | Qwen | $0.08 | $0.28 | 131K | OpenRouter → |
| $0.7 | DeepSeek Flash Latest👁🧠 | DeepSeek | $0.02 | $0.6 | 1.048576M | OpenRouter → |
| $0.7 | Muse Spark 1.3 Contributor👁🧠 | Meta | $0.1 | $0.2 | 1.048576M | OpenRouter → |
| $0.7 | Muse Spark 1.2 Contributor👁🧠 | Meta | $0.1 | $0.2 | 1.048576M | OpenRouter → |
| $0.7 | UI-TARS 7B 👁 | ByteDance | $0.1 | $0.2 | 128K | OpenRouter → |
| $0.7 | Reka Flash 3🧠 | Other | $0.1 | $0.2 | 66K | OpenRouter → |
| $0.7 | Qwen2.5 7B Instruct | Qwen | $0.1 | $0.2 | 33K | OpenRouter → |
| $0.7025 | GLM 4.7 Flash🧠 | Z.ai | $0.0605 | $0.4 | 200K | OpenRouter → |
| $0.7188 | Mistral Small 3.2 24B👁 | Mistral | $0.0938 | $0.25 | 256K | OpenRouter → |
| $0.7425 | Hy3🧠 | Tencent | $0.0825 | $0.33 | 262K | OpenRouter → |
| $0.7875 | Qwen3 235B A22B Instruct 2507 | Qwen | $0.0875 | $0.35 | 262K | OpenRouter → |
| $0.79 | Gemma 4 31B👁🧠 | $0.09 | $0.34 | 262K | OpenRouter → | |
| $0.8 | Step 3.5 Flash🧠 | StepFun | $0.1 | $0.3 | 262K | OpenRouter → |
| $0.8 | Voxtral Small 24B 2507 | Mistral | $0.1 | $0.3 | 33K | OpenRouter → |
| $0.8 | Llama 4 Scout👁 | Meta | $0.1 | $0.3 | 1.31072M | OpenRouter → |
| $0.81 | Solar Pro 4🧠 | Upstage | $0.09 | $0.36 | 524K | OpenRouter → |
| $0.82 | Llama 3.3 70B Instruct | Meta | $0.1 | $0.32 | 131K | OpenRouter → |
| $0.84 | Qwen3 14B🧠 | Qwen | $0.12 | $0.24 | 131K | OpenRouter → |
| $0.85 | Nemotron 3 Super🧠 | NVIDIA | $0.08 | $0.45 | 262K | OpenRouter → |
| $0.85 | Gemma 3 27B👁 | $0.08 | $0.45 | 131K | OpenRouter → | |
| $0.875 | Ternary Bonsai 2 27B👁🧠 | PrismML | $0.075 | $0.5 | 262K | OpenRouter → |
| $0.896 | DeepSeek V4.1 Flash (batch)👁🧠 | DeepSeek | $0.112 | $0.336 | 1.048576M | OpenRouter → |
| $0.9 | Seed-2.0-Mini👁🧠 | ByteDance Seed | $0.1 | $0.4 | 262K | OpenRouter → |
| $0.9 | Gemini 2.5 Flash Lite👁🧠 | $0.1 | $0.4 | 1.048576M | OpenRouter → | |
| $0.9 | GPT-4.1 Nano👁 | OpenAI | $0.1 | $0.4 | 1.047576M | OpenRouter → |
| $0.9 | Ministral 3 8B 2512👁 | Mistral | $0.15 | $0.15 | 262K | OpenRouter → |
| $0.936 | Qwen3 VL 32B Instruct👁 | Qwen | $0.104 | $0.416 | 131K | OpenRouter → |
| $0.98 | MiMo-V2.6-Flash👁🧠 | Xiaomi | $0.14 | $0.28 | 1.05M | OpenRouter → |
| $0.98 | MiMo-V2.5👁🧠 | Xiaomi | $0.14 | $0.28 | 1.05M | OpenRouter → |
| $1 | GPT-6 Luna Pro👁🧠 | OpenAI | $0.1 | $0.5 | 1.05M | OpenRouter → |
| $1 | GPT-6 Luna👁🧠 | OpenAI | $0.1 | $0.5 | 1.05M | OpenRouter → |
| $1 | GPT Luna Latest👁🧠 | OpenAI | $0.1 | $0.5 | 1.05M | OpenRouter → |
| $1.04 | Qwen3 VL 8B Instruct👁 | Qwen | $0.117 | $0.455 | 262K | OpenRouter → |
| $1.04 | Qwen3 8B🧠 | Qwen | $0.117 | $0.455 | 131K | OpenRouter → |
| $1.08 | Llama Guard 4 12B👁 | Meta | $0.18 | $0.18 | 164K | OpenRouter → |
| $1.1 | GPT-5.6 Luna Pro (batch)👁🧠 | OpenAI | $0.1 | $0.6 | 1.05M | OpenRouter → |
| $1.1 | GPT-5.6 Luna (batch)👁🧠 | OpenAI | $0.1 | $0.6 | 1.05M | OpenRouter → |
| $1.1 | Qwen3 30B A3B🧠 | Qwen | $0.12 | $0.5 | 131K | OpenRouter → |
| $1.13 | GPT-5.4 Nano (batch)👁🧠 | OpenAI | $0.1 | $0.625 | 400K | OpenRouter → |
| $1.2 | Nemotron 3.5 Content Safety👁🧠 | NVIDIA | $0.2 | $0.2 | 131K | OpenRouter → |
| $1.2 | Ministral 3 14B 2512👁 | Mistral | $0.2 | $0.2 | 262K | OpenRouter → |
| $1.2 | Codestral 2508 (batch) | Mistral | $0.15 | $0.45 | 256K | OpenRouter → |
| $1.22 | Qwen3.8 Omni Flash👁🧠 | Qwen | $0.15 | $0.47 | 1M | OpenRouter → |
| $1.22 | Qwen3.8 Flash👁🧠 | Qwen | $0.15 | $0.47 | 1M | OpenRouter → |
| $1.25 | GLM 5.3 Flash👁🧠 | Z.ai | $0.15 | $0.5 | 1.048576M | OpenRouter → |
| $1.27 | Hunyuan A13B Instruct🧠 | Tencent | $0.14 | $0.57 | 131K | OpenRouter → |
| $1.35 | Mistral Small 4👁🧠 | Mistral | $0.15 | $0.6 | 262K | OpenRouter → |
| $1.35 | Solar Pro 3🧠 | Upstage | $0.15 | $0.6 | 131K | OpenRouter → |
| $1.35 | Qwen3 VL 30B A3B Instruct👁 | Qwen | $0.15 | $0.6 | 262K | OpenRouter → |
| $1.35 | Command R (08-2024) | Cohere | $0.15 | $0.6 | 128K | OpenRouter → |
| $1.35 | GPT-4o-mini👁 | OpenAI | $0.15 | $0.6 | 128K | OpenRouter → |
| $1.35 | GPT-4o-mini (2024-07-18)👁 | OpenAI | $0.15 | $0.6 | 128K | OpenRouter → |
| $1.37 | DeepSeek V4 Flash Latest🧠 | DeepSeek | $0.0188 | $1.28 | 1.048576M | OpenRouter → |
| $1.37 | DeepSeek V4 Flash 0731🧠 | DeepSeek | $0.0188 | $1.28 | 1.048576M | OpenRouter → |
| $1.38 | Gemini 3.1 Flash Lite (batch)👁🧠 | $0.125 | $0.75 | 1.048576M | OpenRouter → | |
| $1.4 | Qwen3 Coder Next | Qwen | $0.12 | $0.8 | 262K | OpenRouter → |
| $1.46 | DeepSeek V4 Pro 0423🧠 | DeepSeek | $0.2088 | $0.4176 | 1.048576M | OpenRouter → |
| $1.5 | Hy3 preview🧠 | Tencent | $0.18 | $0.6 | 262K | OpenRouter → |
| $1.5 | GLM 4.5 Air🧠 | Z.ai | $0.13 | $0.85 | 131K | OpenRouter → |
| $1.59 | Llama 4 Maverick👁 | Meta | $0.1875 | $0.6525 | 1.048576M | OpenRouter → |
| $1.6 | Saba | Mistral | $0.2 | $0.6 | 33K | OpenRouter → |
| $1.6 | Qwen3 Next 80B A3B Instruct | Qwen | $0.1 | $1.1 | 262K | OpenRouter → |
| $1.63 | GPT-5 Mini (batch)👁🧠 | OpenAI | $0.125 | $1 | 400K | OpenRouter → |
| $1.72 | DeepSeek V4 Flash Vision Exp👁🧠 | DeepSeek | $0.2156 | $0.6468 | 1.048576M | OpenRouter → |
| $1.75 | Qwen3.6 35B A3B👁🧠 | Qwen | $0.15 | $1 | 262K | OpenRouter → |
| $1.75 | Qwen3.5-35B-A3B👁🧠 | Qwen | $0.15 | $1 | 262K | OpenRouter → |
| $1.76 | DeepSeek V3.2 Exp🧠 | DeepSeek | $0.27 | $0.41 | 164K | OpenRouter → |
| $1.8 | GPT-4.1 Mini (batch)👁 | OpenAI | $0.2 | $0.8 | 1.047576M | OpenRouter → |
| $1.82 | DeepSeek V3.2🧠 | DeepSeek | $0.28 | $0.42 | 164K | OpenRouter → |
| $1.89 | MiniMax M2.7🧠 | MiniMax | $0.21 | $0.84 | 205K | OpenRouter → |
| $1.9 | Uncensored | Venice | $0.2 | $0.9 | 128K | OpenRouter → |
| $1.95 | Qwen3 Next 80B A3B Thinking🧠 | Qwen | $0.15 | $1.2 | 262K | OpenRouter → |
| $1.95 | Qwen3 Coder Flash | Qwen | $0.195 | $0.975 | 1M | OpenRouter → |
| $2 | Gemini 3.5 Flash Lite (batch)👁🧠 | $0.15 | $1.25 | 1.048576M | OpenRouter → | |
| $2 | Mercury 2🧠 | Inception | $0.25 | $0.75 | 128K | OpenRouter → |
| $2 | Mistral Large 3 2512 (batch)👁 | Mistral | $0.25 | $0.75 | 262K | OpenRouter → |
| $2 | Cydonia 24B V4.1 | TheDrummer | $0.3 | $0.5 | 131K | OpenRouter → |
| $2 | Mistral Medium 3.1 (batch)👁 | Mistral | $0.2 | $1 | 131K | OpenRouter → |
| $2 | Gemini 2.5 Flash (batch)👁🧠 | $0.15 | $1.25 | 1.048576M | OpenRouter → | |
| $2 | GPT-3.5 Turbo (batch) | OpenAI | $0.25 | $0.75 | 16K | OpenRouter → |
| $2.05 | Trinity Large Thinking🧠 | Arcee AI | $0.25 | $0.8 | 262K | OpenRouter → |
How we rank. We cost your actual workload: input tokens × the input rate plus output tokens × the output rate. That matters because output usually costs 3-6× more than input so a model that looks cheap on input can be dear if you generate a lot. "Blended /1M" assumes a typical 3:1 input:output mix for a quick single-number comparison. All rates are standard USD per million tokens, converted live to your currency, batch APIs are ~50% cheaper and prompt caching can cut cached input up to 90%. Prices move fast; confirm on the provider's site.
Common questions
What's the cheapest LLM API?
Proprietary: Gemini Flash-Lite and GPT-4.1 Nano near $0.10/1M input. Cheapest overall is DeepSeek (~$0.14 in / $0.28 out). The best pick depends on your input:output ratio, use the calculator.
Why do input and output cost different amounts?
Generating output is far more expensive than reading input, so output is usually 3-6× the input price. That's why a single blended price is misleading.
Is Claude or GPT cheaper?
Close at the frontier. Claude Opus ~$5/$25, GPT-5 ~$2.50/$15 standard (much more for Pro). Claude's prompt caching can cut input up to 90%, often making it cheaper in practice.
How do I cut LLM costs?
Use a smaller model where you can (budget models are 20-50× cheaper), cache repeated prompts, use the batch API for non-urgent jobs, and trim prompt and output length.
The other way to pay for tokens
Per-million-token pricing only makes sense against the alternative, which is buying the hardware once and running the model yourself. These track what that hardware costs.
- GPU value rankings performance per dollar, not sticker price
- Used GPU prices what the last generation goes for now