Cheapest reasoning models
Models flagged as reasoning-capable, sorted by input cost, for maths, planning and multi-step debugging.
Current leader: Gemma 4 26b A4b It (Google Gemini). Refreshed 2026-10-28 12:57 UTC.
| # | Model | Provider | Input $/1M | Output $/1M | Cached $/1M | Context |
|---|---|---|---|---|---|---|
| 1 | Gemma 4 26b A4b It | Google Gemini | free | free | — | 262K |
| 2 | Gemma 4 31b It | Google Gemini | free | free | — | 262K |
| 3 | Nemotron 3.5 Lightning 30b A3b | Perplexity | $0.011 | $0.17 | $0.0011 | — |
| 4 | Gemma 4 E4B It | DeepInfra | $0.02 | $0.1 | — | 131K |
| 5 | Qwen3 4b Fp8 | Novita AI | $0.03 | $0.03 | — | 128K |
| 6 | Qwen3 8b Fp8 | Novita AI | $0.035 | $0.138 | — | 128K |
| 7 | GPT Oss 20b | Novita AI | $0.04 | $0.15 | — | 131K |
| 8 | Qwen Turbo | Alibaba DashScope | $0.05 | $0.2 | — | 129K |
| 9 | Qwen Turbo | Alibaba DashScope | $0.05 | $0.2 | — | 1M |
| 10 | Qwen Turbo | Alibaba DashScope | $0.05 | $0.2 | — | 1M |
| 11 | Qwen Turbo | Alibaba DashScope | $0.05 | $0.2 | — | 1M |
| 12 | GPT 5 Nano | OpenAI | $0.05 | $0.4 | $0.005 | 272K |
| 13 | GPT 5 Nano | OpenAI | $0.05 | $0.4 | $0.005 | 272K |
| 14 | GPT 5 Nano | Azure OpenAI | $0.05 | $0.4 | $0.005 | 272K |
| 15 | Nemotron 3 Nano 30B A3B | DeepInfra | $0.05 | $0.2 | $0.025 | 262K |
| 16 | Nemotron Lightning 3p5 30b A3b | Fireworks AI | $0.05 | $0.2 | $0.01 | 262K |
| 17 | GPT Oss 120b | Novita AI | $0.05 | $0.25 | — | 131K |
| 18 | Nemotron 3 Nano 30b A3b | Novita AI | $0.05 | $0.2 | — | 262K |
| 19 | Qwen3 30b A3b Fp8 | Cloudflare Workers AI | $0.051 | $0.335 | — | 33K |
| 20 | GPT 5 Nano | Azure OpenAI | $0.055 | $0.44 | $0.0055 | 272K |
| 21 | GLM 4.7 Flash | DeepInfra | $0.06 | $0.4 | $0.01 | 203K |
| 22 | Ling 3.0 Flash | DeepInfra | $0.06 | $0.18 | $0.012 | 131K |
| 23 | NVIDIA Nemotron 3 Nano 30B A3B | Nebius | $0.06 | $0.24 | — | 262K |
| 24 | Nemotron 3 Nano Omni | Nebius | $0.06 | $0.24 | — | 262K |
| 25 | Nemotron 3 5 Lightning | Nebius | $0.06 | $0.24 | — | 1.0M |
| 26 | Deepseek R1 0528 Qwen3 8b | Novita AI | $0.06 | $0.09 | — | 128K |
| 27 | Ling 3.0 Flash Fast | Novita AI | $0.06 | $0.18 | $0.012 | 262K |
| 28 | Ling 3.0 Flash | Novita AI | $0.06 | $0.18 | $0.012 | 262K |
| 29 | Glm 4.7 Flash | Cloudflare Workers AI | $0.06 | $0.4 | — | 131K |
| 30 | Gemma 4 26B A4B It | DeepInfra | $0.07 | $0.34 | — | 262K |
| 31 | GPT Oss 20b | Fireworks AI | $0.07 | $0.3 | $0.035 | 131K |
| 32 | Ernie 4.5 21B A3b Thinking | Novita AI | $0.07 | $0.28 | — | 131K |
| 33 | Glm 4.7 Flash | Novita AI | $0.07 | $0.4 | $0.01 | 200K |
| 34 | GPT Oss 20b | Groq | $0.075 | $0.3 | $0.037 | 131K |
| 35 | GPT Oss Safeguard 20b | Groq | $0.075 | $0.3 | $0.037 | 131K |
| 36 | NVIDIA Nemotron 3.5 Lightning | DeepInfra | $0.08 | $0.2 | $0.04 | 262K |
| 37 | DeepSeek V4 Flash 0731 | DeepInfra | $0.08 | $0.18 | $0.016 | 1.0M |
| 38 | Openai.gpt Oss 20b 1:0 | Amazon Bedrock | $0.084 | $0.36 | — | 128K |
| 39 | NVIDIA Nemotron 3 Super 120B A12B | DeepInfra | $0.085 | $0.4 | — | 262K |
| 40 | Gemma 4 31B It Turbo | DeepInfra | $0.09 | $0.34 | $0.05 | 262K |
| 41 | DeepSeek V4 Flash | DeepInfra | $0.09 | $0.18 | $0.018 | 1.0M |
| 42 | Qwen3 30b A3b Fp8 | Novita AI | $0.09 | $0.45 | — | 41K |
| 43 | Claude Haiku 5 5 | Anthropic | $0.1 | $0.5 | $0.01 | 1M |
| 44 | Gemini 2.5 Flash Lite | Google Gemini | $0.1 | $0.4 | $0.01 | 1.0M |
| 45 | Muse Spark 1.2 Contributor | Meta | $0.1 | $0.2 | $0.002 | 1.0M |
| 46 | Muse Spark 1.3 Contributor | Meta | $0.1 | $0.2 | $0.002 | 1.0M |
| 47 | GPT 6 Luna | OpenAI | $0.1 | $0.5 | $0.01 | 922K |
| 48 | GPT 6 Luna | Azure OpenAI | $0.1 | $0.5 | $0.01 | 922K |
| 49 | GPT 6 Luna | Azure OpenAI | $0.1 | $0.5 | $0.01 | 922K |
| 50 | GPT Oss 120b | Baseten | $0.1 | $0.5 | $0.1 | 128K |
| 51 | Claude Haiku 5 5 | Amazon Bedrock | $0.1 | $0.5 | $0.01 | 1M |
| 52 | Gemma 4 26b A4b It | Cloudflare Workers AI | $0.1 | $0.3 | — | 256K |
| 53 | Qwen3.6 35B A3B | DeepInfra | $0.1 | $0.95 | — | 262K |
| 54 | Seed 2.0 Mini | DeepInfra | $0.1 | $0.4 | $0.02 | 256K |
| 55 | Cosmos3 Super Reasoner | Nebius | $0.1 | $0.3 | — | 262K |
| 56 | Qwen3 32b Fp8 | Novita AI | $0.1 | $0.45 | — | 41K |
| 57 | Mimo V2 Flash | Novita AI | $0.11 | $0.33 | $0.024 | 262K |
| 58 | Anthropic.claude Haiku 5 5 | Amazon Bedrock | $0.12 | $0.6 | $0.012 | 1M |
| 59 | Deepseek V4 Flash 0731 | Perplexity | $0.13 | $0.26 | $0.028 | — |
| 60 | DeepSeek V4 Flash 0731 | Baseten | $0.13 | $0.26 | $0.028 | 1.0M |
Other use cases
Frequently asked
› Which reasoning model is the cheapest?
Gemma 4 26b A4b It from Google Gemini currently leads this list, at $0/1M input tokens. Rankings are recomputed from the raw dataset on every refresh, so they track price cuts automatically.
› How are these rankings calculated?
Each use case applies a capability filter (context size, vision, tool calling, cache pricing) and then ranks the surviving models by the blended cost that profile actually pays. See the methodology page for the exact formulas.
› Is the cheapest model always the right choice?
No. Check the capability flags and context window, and benchmark quality on your own data — the cheapest model that fails your evals costs more than the second cheapest that passes them.