Models › Cheapest hosted
Cheapest hosted
For calling a hosted model: the cheapest input price among models with two or more providers.
- Fastest first token112msDeepSeek-V4.1-FlashTogether AI via huggingface.co, 2026-10-25
- Cheapest with tool calls$0.010per 1M inQwen3-Embedding-8BDeepInfra via huggingface.co, 2026-10-25
- Most fine-tuned 2025 base2.6kfine-tunesQwen3-4B-Instruct-2507Hugging Face base-model tags, 2026-10-25
Hosted by two or more providers, labelled for the job you pick. Sorted by the cheapest observed input price; $0 rows are a free tier and never count as cheapest. Showing models labelled for tool calls.
The floor
Every view leaves out test fixtures, models whose size didn't parse, and closed models (unless you ask for them). $0 endpoints show as a free tier and never count as cheapest; classifiers and guards never enter a speed sort. How the rank works.
102 models match, first 60 shown. Data observed 2026-10-25 06:54 UTC, Hub snapshot 2026-10-25. Hover a figure for its source and date.
| # | Evidence | |||||||
|---|---|---|---|---|---|---|---|---|
| 1 | Qwen3-Embedding-8BQwen | $0.010 | 2 | 1.0× | — | Apache 2.0 | ||
| 2 | Qwen3-Reranker-0.6BQwen | $0.010 | 2 | 1.0× | — | Apache 2.0 | ||
| 3 | Qwen2.5-Coder-7B-InstructQwen | $0.010 | 2 | 1.0× | 400 ms | Apache 2.0 | ||
| 4 | Qwen3-4B-Thinking-2507Qwen | $0.010 | 2 | 1.0× | 270 ms | Apache 2.0 | ||
| 5 | Qwen2.5-Coder-3B-InstructQwen | $0.010 | 2 | 1.0× | 552 ms | Qwen License (qwen-rese… | ||
| 6 | GLM-5.3-Flashzai-org | $0.018 | 35 | 14.2× | 322 ms | MIT | ||
| 7 | gpt-oss-20bopenai | $0.018 | 13 | 4.2× | 307 ms | Apache 2.0 | ||
| 8 | Qwen3.8-27BQwen | $0.020 | 23 | 49.5× | 205 ms | Apache 2.0 | ||
| 9 | DeepSeek-V4.1-Flashdeepseek-ai | $0.020 | 34 | 22.7× | 112 ms | MIT | ||
| 10 | Llama-3.1-8B-Instructmeta-llama | $0.020 | 4 | 3.0× | 269 ms | Llama 3.1 Community Lic… | ||
| 11 | Ling-3.0-flashinclusionAI | $0.021 | 2 | 2.9× | 806 ms | MIT | ||
| 12 | gpt-oss-120bopenai | $0.030 | 27 | 11.8× | 172 ms | Apache 2.0 | ||
| 13 | DeepSeek-V4-Flashdeepseek-ai | $0.030 | 16 | 16.4× | 464 ms | MIT | ||
| 14 | GLM-5.2zai-org | $0.034 | 28 | 60.2× | 224 ms | MIT | ||
| 15 | gemma-3-4b-itgoogle | $0.040 | 3 | 1.3× | 327 ms | Gemma Terms of Use | ||
| 16 | gemma-3-12b-itgoogle | $0.040 | 3 | 1.9× | 302 ms | Gemma Terms of Use | ||
| 17 | gemma-4-26B-A4B-itgoogle | $0.042 | 16 | 6.8× | 363 ms | Apache 2.0 | ||
| 18 | gemma-4-26B-A4Bgoogle | $0.042 | 15 | 3.6× | — | Apache 2.0 | ||
| 19 | DeepSeek-V4-Flash-0731deepseek-ai | $0.042 | 24 | 10.4× | 227 ms | MIT | ||
| 20 | Qwen3.6-35B-A3BQwen | $0.050 | 12 | 5.7× | 265 ms | Apache 2.0 | ||
| 21 | Llama-3.2-3B-Instructmeta-llama | $0.050 | 3 | 1.0× | — | Llama 3.2 Community Lic… | ||
| 22 | DeepSeek-R1-Distill-Llama-8Bdeepseek-ai | $0.050 | 2 | 1.0× | 311 ms | MIT | ||
| 23 | phi-4microsoft | $0.056 | 2 | 1.0× | 520 ms | MIT | ||
| 24 | GLM-4.7-Flashzai-org | $0.060 | 5 | 1.2× | 713 ms | MIT | ||
| 25 | Qwen2.5-Coder-32B-InstructQwen | $0.060 | 3 | 11.0× | 463 ms | Apache 2.0 | ||
| 26 | granite-4.2-8bibm-granite | $0.060 | 2 | 1.7× | 324 ms | Apache 2.0 | ||
| 27 | Qwen3.5-35B-A3BQwen | $0.065 | 8 | 3.8× | 372 ms | Apache 2.0 | ||
| 28 | Qwen3-8BQwen | $0.070 | 2 | 1.0× | 437 ms | Apache 2.0 | ||
| 29 | Qwen3-14BQwen | $0.070 | 4 | 1.7× | 154 ms | Apache 2.0 | ||
| 30 | Qwen3-Coder-30B-A3B-InstructQwen | $0.070 | 3 | 2.1× | — | Apache 2.0 | ||
| 31 | Qwen3-Next-80B-A3B-InstructQwen | $0.072 | 4 | 2.1× | 545 ms | Apache 2.0 | ||
| 32 | GLM-5.3zai-org | $0.075 | 38 | 28.0× | 329 ms | GLM License (glm-5.3) | ||
| 33 | Qwen3.5-9BQwen | $0.080 | 7 | 2.1× | 279 ms | Apache 2.0 | ||
| 34 | Qwen3-32BQwen | $0.080 | 5 | 1.8× | 421 ms | Apache 2.0 | ||
| 35 | gemma-3-27b-itgoogle | $0.080 | 6 | 1.5× | 1395 ms | Gemma Terms of Use | ||
| 36 | Llama-4-Scout-17B-16E-Instructmeta-llama | $0.080 | 4 | 3.1× | 369 ms | Llama Community License… | ||
| 37 | Hy3tencent | $0.083 | 7 | 2.4× | 674 ms | Apache 2.0 | ||
| 38 | Qwen3-235B-A22B-Instruct-2507Qwen | $0.087 | 7 | 9.8× | 200 ms | Apache 2.0 | ||
| 39 | Qwen3-30B-A3B-Instruct-2507Qwen | $0.090 | 4 | 1.1× | — | Apache 2.0 | ||
| 40 | MiMo-V2.6-Flash-RLXiaomiMiMo | $0.091 | 10 | 1.9× | 4234 ms | MIT | ||
| 41 | gemma-4-31B-itgoogle | $0.10 | 14 | 7.5× | 301 ms | Apache 2.0 | ||
| 42 | Llama-3.3-70B-Instructmeta-llama | $0.10 | 12 | 10.3× | 278 ms | Llama 3.3 Community Lic… | ||
| 43 | Qwen2.5-7B-InstructQwen | $0.10 | 2 | 1.0× | — | Apache 2.0 | ||
| 44 | Step-3.5-Flashstepfun-ai | $0.10 | 2 | 1.0× | — | Apache 2.0 | ||
| 45 | UI-TARS-1.5-7BByteDance-Seed | $0.10 | 2 | 1.0× | — | Apache 2.0 | ||
| 46 | Qwen2.5-7BQwen | $0.10 | 2 | 1.0× | — | Apache 2.0 | ||
| 47 | gemma-4-31Bgoogle | $0.10 | 13 | 7.5× | — | Apache 2.0 | ||
| 48 | Apertus-8B-Instruct-2509swiss-ai | $0.10 | 2 | 1.0× | 4683 ms | Apache 2.0 | ||
| 49 | DeepSeek-V4-Pro-0813deepseek-ai | $0.12 | 22 | 13.9× | 417 ms | MIT | ||
| 50 | MiMo-V2.5XiaomiMiMo | $0.12 | 3 | 3.4× | 2231 ms | MIT | ||
| 51 | Qwen3-Coder-NextQwen | $0.12 | 2 | 1.0× | — | Apache 2.0 | ||
| 52 | GLM-4.5-Airzai-org | $0.13 | 3 | 1.5× | 1104 ms | MIT | ||
| 53 | Qwen3.8-Flash-NextQwen | $0.15 | 2 | 1.0× | — | Qwen License (qwen-comm… | ||
| 54 | Qwen3-VL-30B-A3B-InstructQwen | $0.15 | 4 | 1.9× | 335 ms | Apache 2.0 | ||
| 55 | DeepSeek-R1-Distill-Qwen-7Bdeepseek-ai | $0.15 | 2 | 1.0× | 437 ms | MIT | ||
| 56 | MiniMax-M3.1-FlashMiniMaxAI | $0.15 | 2 | 1.0× | 1962 ms | MiniMax Model License (… | ||
| 57 | Qwen3.5-27BQwen | $0.20 | 7 | 1.5× | 462 ms | Apache 2.0 | ||
| 58 | Qwen3.6-27BQwen | $0.20 | 9 | 2.3× | 250 ms | Apache 2.0 | ||
| 59 | Qwen3-235B-A22BQwen | $0.20 | 2 | 1.0× | 578 ms | Apache 2.0 | ||
| 60 | DeepSeek-V4.1-Flash-1017deepseek-ai | $0.20 | 5 | 1.5× | 171 ms | MIT |
Evidence marks, left to right: coding, tools, reasoning, long docs, vision, chat. measured on a third-party board, author-reported, structural, none yet. Each mark links to its job from the model page. INT4 is an estimate of the weights alone (method).