Fastest AI Models
Models ranked by measured p50 latency. Lower is better. All data from real ANT routing requests.
Updated October 14, 2026
| Rank | Model | Provider | p50 Latency | p95 Latency | Score | Success Rate |
|---|---|---|---|---|---|---|
| 1 | command-a-03-2025 | semantic-cache | 0ms | 0ms | 77.0 | 10000.0% |
| 2 | deepseek-ai/DeepSeek-V4.1-Flash | semantic-cache | 0ms | 0ms | 77.0 | 10000.0% |
| 3 | gemini-2.5-flash | semantic-cache | 0ms | 0ms | 77.0 | 10000.0% |
| 4 | Qwen: Qwen3 Coder Next | semantic-cache | 0ms | 0ms | 77.0 | 10000.0% |
| 5 | meta-llama/Llama-4-Scout-17B-16E-Instruct | semantic-cache | 0ms | 0ms | 77.0 | 10000.0% |
| 6 | deepseek/deepseek-r1-turbo | semantic-cache | 0ms | 0ms | 61.0 | 10000.0% |
| 7 | deepseek-ai/DeepSeek-V3-0324 | semantic-cache | 0ms | 0ms | 61.0 | 10000.0% |
| 8 | deepseek/deepseek-v3-0324 | semantic-cache | 0ms | 0ms | 58.0 | 10000.0% |
| 9 | DeepSeek: DeepSeek V4 Flash 0731 | semantic-cache | 0ms | 0ms | 58.0 | 10000.0% |
| 10 | TheDrummer/Rocinante-12B-v1.1 | semantic-cache | 0ms | 0ms | 58.0 | 10000.0% |
| 11 | Qwen: Qwen3.7 Flash | semantic-cache | 0ms | 0ms | 57.0 | 10000.0% |
| 12 | GPT-4o Mini | semantic-cache | 0ms | 0ms | 57.0 | 10000.0% |
| 13 | NVIDIA: Nemotron 3 Nano 30B A3B | semantic-cache | 0ms | 0ms | 55.0 | 10000.0% |
| 14 | OpenAI: gpt-oss-20b | semantic-cache | 0ms | 0ms | 55.0 | 10000.0% |
| 15 | llama-3.3-70b-instruct | semantic-cache | 0ms | 0ms | 55.0 | 10000.0% |
| 16 | Silicone-Moss/CrucibleLab-L3.3-70B-Loki-V2.0-Heretic-Uncensored | semantic-cache | 0ms | 0ms | 55.0 | 10000.0% |
| 17 | DeepHat/DeepHat-V1-7B | semantic-cache | 0ms | 0ms | 55.0 | 10000.0% |
| 18 | miss-t/Qwen3.8-27B-AEON-ULTIMATE-UNCENSORED-BF16 | semantic-cache | 0ms | 0ms | 55.0 | 10000.0% |
| 19 | OpenAI: gpt-oss-120b | semantic-cache | 0ms | 0ms | 55.0 | 10000.0% |
| 20 | mistral-small-latest | semantic-cache | 0ms | 0ms | 55.0 | 10000.0% |
| 21 | Vortex5/Shadow-Siren-26B-A4B | semantic-cache | 0ms | 0ms | 55.0 | 10000.0% |
| 22 | huihui-ai/Huihui-Mistral-Small-3.2-24B-Instruct-2506-abliterated | semantic-cache | 0ms | 0ms | 55.0 | 10000.0% |
| 23 | deepseek-v4-flash | semantic-cache | 0ms | 0ms | 55.0 | 10000.0% |
| 24 | cloudbjorn/Qwen3.8-27B-Yes-Man-uncensored | semantic-cache | 0ms | 0ms | 55.0 | 10000.0% |
| 25 | TheDrummer/Cydonia-24B-v4.3 | semantic-cache | 0ms | 0ms | 55.0 | 10000.0% |
| 26 | huihui-ai/Qwen2.5-72B-Instruct-abliterated | semantic-cache | 0ms | 0ms | 55.0 | 10000.0% |
| 27 | o3-mini | semantic-cache | 0ms | 0ms | 55.0 | 10000.0% |
| 28 | huihui-ai/Llama-3.3-70B-Instruct-abliterated | semantic-cache | 0ms | 0ms | 55.0 | 10000.0% |
| 29 | philbert440/Qwen3.8-27B-Uncensored-Cyber | semantic-cache | 0ms | 0ms | 55.0 | 10000.0% |
| 30 | Qwen: Qwen3 Next 80B A3B Thinking | semantic-cache | 0ms | 0ms | 55.0 | 10000.0% |
| 31 | Qwen: Qwen3.5-9B | semantic-cache | 0ms | 0ms | 55.0 | 10000.0% |
| 32 | GPT-4.1 Nano | semantic-cache | 0ms | 0ms | 55.0 | 10000.0% |
| 33 | Qwen2.5 72B Instruct Abliterated | unknown | 1ms | 1ms | 46.0 | 0.0% |
| 34 | ant:best | unknown | 2ms | 2ms | 46.0 | 0.0% |
| 35 | ant:vision | unknown | 2ms | 2ms | 46.0 | 0.0% |
| 36 | ant:coder | unknown | 2ms | 2ms | 46.0 | 0.0% |
| 37 | ant:reasoning-fast | unknown | 2ms | 2ms | 46.0 | 0.0% |
| 38 | Llama 3.3 70B Instruct Abliterated | unknown | 2ms | 2ms | 46.0 | 0.0% |
| 39 | Huihui Mistral Small 3.2 24B Instruct 2506 Abliterated | unknown | 2ms | 2ms | 46.0 | 0.0% |
| 40 | Qwen3.8 27B AEON ULTIMATE UNCENSORED BF16 | unknown | 2ms | 2ms | 46.0 | 0.0% |
| 41 | Qwen3.8 27B Yes Man Uncensored | unknown | 2ms | 2ms | 46.0 | 0.0% |
| 42 | Qwen3.8 27B Uncensored Cyber | unknown | 2ms | 2ms | 46.0 | 0.0% |
| 43 | ant:reasoning-premium | unknown | 2ms | 2ms | 46.0 | 0.0% |
| 44 | ant:fast | unknown | 3ms | 3ms | 46.0 | 0.0% |
| 45 | ant:balanced | unknown | 3ms | 3ms | 46.0 | 0.0% |
| 46 | SpaceXAI: Grok 4.20 Multi-Agent | unknown | 3ms | 3ms | 35.0 | 0.0% |
| 47 | DeepHat V1 7B | unknown | 6ms | 6ms | 40.0 | 0.0% |
| 48 | kimi-k3 | unknown | 6ms | 6ms | 35.0 | 0.0% |
| 49 | Command A (03-2025) | unknown | 7ms | 7ms | 35.0 | 0.0% |
| 50 | ant:reasoning | unknown | 7ms | 7ms | 35.0 | 0.0% |