Models › Low latency
Low latency
For realtime jobs: chat models with a measured first-token time, fastest first.
- 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
General chat models with a measured first-token time, fastest first. Classifiers and safety models are left out so a guard model can't top a chat list.
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.
Kind: All kinds · LLMs · Embeddings · Decision models · Closed
88 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 | DeepSeek-V4.1-Flashdeepseek-ai | 112 ms | 309 | $0.020 | 34 | MIT | |
| 2 | Qwen3-14BQwen | 154 ms | 104 | $0.070 | 4 | Apache 2.0 | |
| 3 | gpt-oss-120bopenai | 172 ms | 407 | $0.030 | 27 | Apache 2.0 | |
| 4 | Qwen3-235B-A22B-Instruct-2507Qwen | 200 ms | 35 | $0.087 | 7 | Apache 2.0 | |
| 5 | Qwen3.8-27BQwen | 205 ms | 995 | $0.020 | 23 | Apache 2.0 | |
| 6 | Inklingthinkingmachines | 208 ms | 142 | $0.95 | 4 | Apache 2.0 | |
| 7 | GLM-5.2zai-org | 224 ms | 201 | $0.034 | 28 | MIT | |
| 8 | c4ai-command-r7b-12-2024CohereLabs | 229 ms | 78 | — | 1 | CC BY-NC 4.0 | |
| 9 | Qwen3-30B-A3BQwen | 230 ms | 134 | $0.12 | 1 | Apache 2.0 | |
| 10 | Qwen3.5-122B-A10BQwen | 232 ms | 112 | $0.26 | 5 | Apache 2.0 | |
| 11 | Hermes-3-Llama-3.1-70BNousResearch | 235 ms | 28 | $0.70 | 2 | Llama 3 Community Licen… | |
| 12 | Qwen3.6-27BQwen | 250 ms | 57 | $0.20 | 9 | Apache 2.0 | |
| 13 | Qwen3-235B-A22B-Thinking-2507Qwen | 264 ms | 145 | $0.23 | 3 | Apache 2.0 | |
| 14 | Qwen3.6-35B-A3BQwen | 265 ms | 106 | $0.050 | 12 | Apache 2.0 | |
| 15 | Muse-Glimmer-30Bmeta-models | 268 ms | 98 | $0.30 | 3 | Apache 2.0 | |
| 16 | Llama-3.1-8B-Instructmeta-llama | 269 ms | 152 | $0.020 | 4 | Llama 3.1 Community Lic… | |
| 17 | Qwen3-4B-Thinking-2507Qwen | 270 ms | 161 | $0.010 | 2 | Apache 2.0 | |
| 18 | Llama-3.3-70B-Instructmeta-llama | 278 ms | 81 | $0.10 | 12 | Llama 3.3 Community Lic… | |
| 19 | Qwen3.5-9BQwen | 279 ms | 93 | $0.080 | 7 | Apache 2.0 | |
| 20 | Qwen3.8-2.4T-A95BQwen | 284 ms | 118 | $2.00 | 7 | Qwen License (qwen3.8-m… | |
| 21 | gemma-4-31B-itgoogle | 301 ms | 56 | $0.10 | 14 | Apache 2.0 | |
| 22 | gemma-3-12b-itgoogle | 302 ms | 72 | $0.040 | 3 | Gemma Terms of Use | |
| 23 | gpt-oss-20bopenai | 307 ms | 721 | $0.018 | 13 | Apache 2.0 | |
| 24 | DeepSeek-R1-Distill-Llama-8Bdeepseek-ai | 311 ms | 137 | $0.050 | 2 | MIT | |
| 25 | GLM-5.3-Flashzai-org | 322 ms | 130 | $0.018 | 35 | MIT | |
| 26 | granite-4.2-8bibm-granite | 324 ms | 34 | $0.060 | 2 | Apache 2.0 | |
| 27 | gemma-3-4b-itgoogle | 327 ms | 91 | $0.040 | 3 | Gemma Terms of Use | |
| 28 | GLM-5.3zai-org | 329 ms | 95 | $0.075 | 38 | GLM License (glm-5.3) | |
| 29 | GLM-4.6zai-org | 334 ms | 38 | $0.40 | 5 | MIT | |
| 30 | Qwen3-VL-30B-A3B-InstructQwen | 335 ms | 38 | $0.15 | 4 | Apache 2.0 | |
| 31 | MiniMax-M3MiniMaxAI | 336 ms | 106 | $0.22 | 14 | MiniMax Model License (… | |
| 32 | GLM-5zai-org | 342 ms | 81 | $0.60 | 10 | MIT | |
| 33 | DeepSeek-V3-0324deepseek-ai | 360 ms | 85 | $0.24 | 4 | MIT | |
| 34 | gemma-4-26B-A4B-itgoogle | 363 ms | 180 | $0.042 | 16 | Apache 2.0 | |
| 35 | Llama-4-Scout-17B-16E-Instructmeta-llama | 369 ms | 70 | $0.080 | 4 | Llama Community License… | |
| 36 | Qwen3.5-35B-A3BQwen | 372 ms | 129 | $0.065 | 8 | Apache 2.0 | |
| 37 | Qwen3-4B-Instruct-2507Qwen | 386 ms | 24 | $0.010 | 1 | Apache 2.0 | |
| 38 | Qwen3.5-397B-A17BQwen | 396 ms | 97 | $0.39 | 13 | Apache 2.0 | |
| 39 | Qwen2.5-Coder-7B-InstructQwen | 400 ms | 143 | $0.010 | 2 | Apache 2.0 | |
| 40 | Kimi-K2.6moonshotai | 414 ms | 41 | $0.45 | 18 | Kimi (Modified MIT) (mo… | |
| 41 | DeepSeek-V4-Pro-0813deepseek-ai | 417 ms | 119 | $0.12 | 22 | MIT | |
| 42 | Qwen3-32BQwen | 421 ms | 46 | $0.080 | 5 | Apache 2.0 | |
| 43 | Qwen3-8BQwen | 437 ms | 87 | $0.070 | 2 | Apache 2.0 | |
| 44 | DeepSeek-R1-Distill-Qwen-7Bdeepseek-ai | 437 ms | 145 | $0.15 | 2 | MIT | |
| 45 | MiMo-V2.5-ProXiaomiMiMo | 447 ms | 2 | $0.30 | 4 | MIT | |
| 46 | Kimi-K3moonshotai | 451 ms | 62 | $0.23 | 22 | Kimi K3 (Modified MIT)… | |
| 47 | Qwen3.5-27BQwen | 462 ms | 68 | $0.20 | 7 | Apache 2.0 | |
| 48 | Qwen2.5-Coder-32B-InstructQwen | 463 ms | 39 | $0.060 | 3 | Apache 2.0 | |
| 49 | DeepSeek-V4-Flashdeepseek-ai | 464 ms | 33 | $0.030 | 16 | MIT | |
| 50 | DeepSeek-V4-Flash-Vision-Expdeepseek-ai | 470 ms | 69 | $0.22 | 3 | MIT | |
| 51 | DeepSeek-V4-Prodeepseek-ai | 477 ms | 57 | $0.21 | 16 | MIT | |
| 52 | DeepSeek-R1-Distill-Qwen-14Bdeepseek-ai | 483 ms | 86 | $0.20 | 2 | MIT | |
| 53 | Kimi-K2.5moonshotai | 494 ms | 95 | $0.45 | 7 | Kimi (Modified MIT) (mo… | |
| 54 | GLM-5.1zai-org | 500 ms | 75 | $0.96 | 15 | MIT | |
| 55 | Inkling-Smallthinkingmachines | 513 ms | 143 | $0.45 | 2 | Apache 2.0 | |
| 56 | phi-4microsoft | 520 ms | 91 | $0.056 | 2 | MIT | |
| 57 | L3-8B-Stheno-v3.2Sao10K | 525 ms | 110 | $0.050 | 2 | CC BY-NC 4.0 | |
| 58 | Qwen3-Next-80B-A3B-InstructQwen | 545 ms | 121 | $0.072 | 4 | Apache 2.0 | |
| 59 | Qwen2.5-Coder-3B-InstructQwen | 552 ms | 169 | $0.010 | 2 | Qwen License (qwen-rese… | |
| 60 | Qwen3-235B-A22BQwen | 578 ms | 19 | $0.20 | 2 | Apache 2.0 |
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).