Benchmarks
Leaderboard
Every active text model we can reach, with how long it takes to answer and how often it wins a head-to-head. Nothing is hidden: a model we have not measured still gets a row, and the row says why.
- Typical
- How long a normal answer takes. Half of the calls came back faster than this, half slower — so it is what you should expect on an ordinary request.
- Slow case
- How bad it gets on an off day. Roughly one call in twenty is slower than this. A model with a low typical time but a high slow case is fast until it is not.
- Win rate
- Out of 100 head-to-head comparisons against an average model on this board, how many this one wins. 50 is average, higher is better. It is worked out from every judged comparison, not from a single score.
108 of 112 models fully measured · 0 timed once · 0 awaiting a first run · 4 we cannot measure.
Filter:
| # | |||||||
|---|---|---|---|---|---|---|---|
| 1 | Qwen2.5-VL-72B-InstructB | OVH AI Endpoints (GRA) | B | 143 ms | 145 ms | 53% | 2026-11-12 |
| 2 | NVIDIA Nemotron Super 49B v1.5A | OpenRouter | A | 173 ms | 199 ms | not compared | 2026-07-17 |
| 3 | Nous Hermes 3 70BA | OpenRouter | A | 212 ms | 883 ms | not compared | 2026-11-02 |
| 4 | DeepSeek v3.2A | OpenRouter | A | 219 ms | 680 ms | not compared | 2026-11-12 |
| 5 | Llama 3.3 70B InstructA | OpenRouter | A | 243 ms | 252 ms | not compared | 2026-11-12 |
| 6 | Llama 4 ScoutA | OpenRouter | A | 255 ms | 261 ms | not compared | 2026-11-12 |
| 7 | gpt-oss-20bC | OVH AI Endpoints (GRA) | C | 258 ms | 453 ms | 36% | 2026-11-12 |
| 8 | Gemini 2.5 Flash-LiteB | Google Gemini | B | 369 ms | 408 ms | 50% | 2026-11-12 |
| 9 | gpt-4.1-mini-2025-04-14C | OpenAI | C | 406 ms | 438 ms | 67% | 2026-11-12 |
| 10 | gpt-4.1-miniC | OpenAI | C | 417 ms | 540 ms | 71% | 2026-11-12 |
| 11 | Claude Haiku 4.5A | Anthropic | A | 436 ms | 441 ms | 53% | 2026-11-12 |
| 12 | Llama 4 MaverickA | OpenRouter | A | 451 ms | 457 ms | not compared | 2026-11-12 |
| 13 | Meta-Llama-3_3-70B-InstructB | OVH AI Endpoints (GRA) | B | 451 ms | 3.7 s | 53% | 2026-11-12 |
| 14 | Qwen 2.5 VL 72B InstructA | OpenRouter | A | 459 ms | 540 ms | not compared | 2026-11-12 |
| 15 | Grok 4.20 Non-Reasoning (0309)B | xAI | B | 472 ms | 485 ms | not compared | 2026-11-12 |
| 16 | Mistral-Small-3.2-24B-Instruct-2506B | OVH AI Endpoints (GRA) | B | 480 ms | 3.8 s | 52% | 2026-11-12 |
| 17 | gpt-5.4-nano-2026-03-17A | OpenAI | A | 488 ms | 595 ms | 58% | 2026-11-12 |
| 18 | gpt-5.4-nanoC | OpenAI | C | 495 ms | 620 ms | 54% | 2026-11-12 |
| 19 | gpt-4.1B | OpenAI | B | 496 ms | 731 ms | 66% | 2026-11-12 |
| 20 | Gemini Flash-Lite LatestC | Google Gemini | C | 497 ms | 556 ms | 63% | 2026-11-12 |
| 21 | gpt-4o-mini-2024-07-18C | OpenAI | C | 511 ms | 611 ms | 54% | 2026-11-12 |
| 22 | gpt-4oC | OpenAI | C | 512 ms | 537 ms | 64% | 2026-11-12 |
| 23 | gpt-4.1-2025-04-14C | OpenAI | C | 515 ms | 522 ms | 70% | 2026-11-12 |
| 24 | Gemini 3.5 Flash-LiteB | Google Gemini | B | 521 ms | 793 ms | 68% | 2026-11-12 |
| 25 | gpt-4o-miniC | OpenAI | C | 525 ms | 582 ms | 58% | 2026-11-12 |
| 26 | gpt-4o-2024-11-20C | OpenAI | C | 529 ms | 902 ms | 60% | 2026-11-12 |
| 27 | o3-2025-04-16B | OpenAI | B | 545 ms | 564 ms | 62% | 2026-11-12 |
| 28 | Qwen3.5-9BB | OVH AI Endpoints (GRA) | B | 546 ms | 564 ms | 22% | 2026-11-12 |
| 29 | MiniMax M2.5A | OpenRouter | A | 550 ms | 587 ms | not compared | 2026-11-12 |
| 30 | gpt-5-nano-2025-08-07B | OpenAI | B | 554 ms | 590 ms | 32% | 2026-11-12 |
| 31 | Gemini 3.1 Flash LiteB | Google Gemini | B | 575 ms | 624 ms | 61% | 2026-11-12 |
| 32 | gpt-5-nanoC | OpenAI | C | 600 ms | 612 ms | 26% | 2026-11-12 |
| 33 | gpt-4o-2024-08-06C | OpenAI | C | 604 ms | 677 ms | 62% | 2026-11-12 |
| 34 | gpt-oss-120bC | OVH AI Endpoints (GRA) | C | 606 ms | 1.1 s | 60% | 2026-11-12 |
| 35 | gpt-5.2-2025-12-11B | OpenAI | B | 633 ms | 723 ms | 68% | 2026-11-12 |
| 36 | gpt-5.4-miniA | OpenAI | A | 654 ms | 843 ms | 65% | 2026-11-12 |
| 37 | DeepSeek v4 ProA | OpenRouter | A | 657 ms | 1.8 s | not compared | 2026-11-12 |
| 38 | gpt-5.6-terra | Azure OpenAI (EU - Sweden) | unrated | 671 ms | 1.1 s | 71% | 2026-11-12 |
| 39 | Gemini 2.5 FlashA | Google Gemini | A | 672 ms | 8.2 s | 15% | 2026-11-12 |
| 40 | Gemini 3 Flash PreviewC | Google Gemini | C | 677 ms | 695 ms | 45% | 2026-11-12 |
| 41 | gpt-5.1-2025-11-13B | OpenAI | B | 681 ms | 891 ms | 70% | 2026-11-12 |
| 42 | Claude Sonnet 5A | Anthropic | A | 685 ms | 718 ms | 59% | 2026-11-12 |
| 43 | gpt-5.1B | OpenAI | B | 686 ms | 699 ms | 72% | 2026-11-12 |
| 44 | gpt-5.6-terraB | OpenAI | B | 686 ms | 702 ms | 71% | 2026-11-12 |
| 45 | gpt-5.4A | OpenAI | A | 688 ms | 763 ms | 66% | 2026-11-12 |
| 46 | gpt-5.6-lunaB | OpenAI | B | 689 ms | 794 ms | 66% | 2026-11-12 |
| 47 | gpt-5.4-mini-2026-03-17A | OpenAI | A | 702 ms | 944 ms | 66% | 2026-11-12 |
| 48 | gpt-6-lunaB | OpenAI | B | 702 ms | 768 ms | 65% | 2026-11-12 |
| 49 | gpt-5.4-2026-03-05B | OpenAI | B | 706 ms | 882 ms | 69% | 2026-11-12 |
| 50 | gpt-6-solA | OpenAI | A | 709 ms | 883 ms | 53% | 2026-11-12 |
| 51 | Claude Opus 4.5B | Anthropic | B | 712 ms | 779 ms | 63% | 2026-11-12 |
| 52 | Gemini 3.5 FlashA | Google Gemini | A | 732 ms | 1.0 s | 26% | 2026-11-12 |
| 53 | Claude Sonnet 4.5B | Anthropic | B | 748 ms | 764 ms | 61% | 2026-11-12 |
| 54 | o3C | OpenAI | C | 761 ms | 935 ms | 64% | 2026-11-12 |
| 55 | Qwen 3.6 PlusA | OpenRouter | A | 784 ms | 1.1 s | not compared | 2026-11-12 |
| 56 | Claude Opus 5A | Anthropic | A | 790 ms | 809 ms | 52% | 2026-11-12 |
| 57 | gpt-5-miniC | OpenAI | C | 794 ms | 1.1 s | 57% | 2026-11-12 |
| 58 | Claude Opus 4.7B | Anthropic | B | 822 ms | 1.1 s | 65% | 2026-11-12 |
| 59 | gpt-5-2025-08-07B | OpenAI | B | 842 ms | 942 ms | 44% | 2026-11-12 |
| 60 | gpt-5.2B | OpenAI | B | 843 ms | 1.0 s | 69% | 2026-11-12 |
| 61 | gpt-5C | OpenAI | C | 856 ms | 954 ms | 44% | 2026-11-12 |
| 62 | Claude Sonnet 5.5A | Anthropic | A | 859 ms | 994 ms | not compared | 2026-11-12 |
| 63 | gpt-5-mini-2025-08-07B | OpenAI | B | 878 ms | 915 ms | 53% | 2026-11-12 |
| 64 | Claude Opus 5.5A | Anthropic | A | 896 ms | 1.0 s | not compared | 2026-11-12 |
| 65 | Cohere Command-AA | OpenRouter | A | 913 ms | 5.9 s | not compared | 2026-11-12 |
| 66 | Grok Build 0.1B | xAI | B | 918 ms | 1.3 s | not compared | 2026-11-12 |
| 67 | Qwen 3.7 MaxA | OpenRouter | A | 958 ms | 1.2 s | not compared | 2026-11-12 |
| 68 | Gemini 3.6 FlashB | Google Gemini | B | 967 ms | 1.0 s | 14% | 2026-11-12 |
| 69 | Claude Fable 5.1A | Anthropic | A | 977 ms | 1.2 s | not compared | 2026-11-12 |
| 70 | Claude Opus 4.8A | Anthropic | A | 1.0 s | 1.1 s | 60% | 2026-11-12 |
| 71 | gpt-6-lunaB | Azure OpenAI (EU - Sweden) | B | 1.1 s | 1.1 s | 65% | 2026-11-12 |
| 72 | Grok 4.5B | xAI | B | 1.1 s | 1.1 s | not compared | 2026-11-12 |
| 73 | Qwen3.7 MaxA | Alibaba Cloud Qwen (DashScope Intl) | A | 1.1 s | 1.8 s | not compared | 2026-11-12 |
| 74 | gpt-5.5C | OpenAI | C | 1.1 s | 1.3 s | 71% | 2026-11-12 |
| 75 | Gemini 3.1 Pro Preview Custom ToolsC | Google Gemini | C | 1.3 s | 1.8 s | 23% | 2026-05-31 |
| 76 | Grok 4.20 Reasoning (0309)B | xAI | B | 1.3 s | 1.4 s | not compared | 2026-11-12 |
| 77 | Qwen3.7 PlusB | Alibaba Cloud Qwen (DashScope Intl) | B | 1.3 s | 1.4 s | not compared | 2026-11-12 |
| 78 | GLM-4.6V (vision)A | Z.ai (GLM / Zhipu) | A | 1.3 s | 1.9 s | 29% | 2026-11-12 |
| 79 | GLM-4.5V (vision)A | Z.ai (GLM / Zhipu) | A | 1.4 s | 2.5 s | 30% | 2026-11-12 |
| 80 | gpt-5.5-2026-04-23A | OpenAI | A | 1.4 s | 1.8 s | 71% | 2026-11-12 |
| 81 | Grok 4.7A | xAI | A | 1.4 s | 1.6 s | not compared | 2026-11-12 |
| 82 | gpt-5-search-apiC | OpenAI | C | 1.4 s | 1.4 s | 62% | 2026-11-12 |
| 83 | gpt-6-astraA | OpenAI | A | 1.4 s | 1.8 s | 66% | 2026-11-12 |
| 84 | gpt-5-search-api-2025-10-14B | OpenAI | B | 1.5 s | 1.7 s | 62% | 2026-11-12 |
| 85 | Claude Opus 4.6B | Anthropic | B | 1.6 s | 1.6 s | 68% | 2026-11-12 |
| 86 | Gemini Pro LatestC | Google Gemini | C | 1.6 s | 2.0 s | 19% | 2026-11-12 |
| 87 | gpt-6-astraA | Azure OpenAI (EU - Sweden) | A | 1.6 s | 1.8 s | 66% | 2026-11-12 |
| 88 | Gemini 2.5 ProA | Google Gemini | A | 1.7 s | 2.0 s | 19% | 2026-11-12 |
| 89 | gpt-6.1-solA | Azure OpenAI (EU - Sweden) | A | 1.7 s | 1.8 s | 65% | 2026-11-12 |
| 90 | Gemini 3.8 FlashA | Google Gemini | A | 1.8 s | 2.2 s | 35% | 2026-11-12 |
| 91 | Gemini 3.1 Pro PreviewC | Google Gemini | C | 1.9 s | 1.9 s | 22% | 2026-11-12 |
| 92 | Claude Fable 5A | Anthropic | A | 2.2 s | 2.3 s | 62% | 2026-11-12 |
| 93 | Gemini Flash LatestB | Google Gemini | B | 2.2 s | 2.2 s | 34% | 2026-11-12 |
| 94 | Gemini 3.7 FlashB | Google Gemini | B | 2.3 s | 5.2 s | 46% | 2026-11-12 |
| 95 | Grok 4.3B | xAI | B | 2.3 s | 3.7 s | not compared | 2026-11-12 |
| 96 | GLM-4.5 AirB | Z.ai (GLM / Zhipu) | B | 2.5 s | 3.9 s | 36% | 2026-11-12 |
| 97 | GLM-5.2A | Z.ai (GLM / Zhipu) | A | 2.6 s | 3.2 s | 26% | 2026-11-12 |
| 98 | Grok 4.6B | xAI | B | 3.0 s | 3.4 s | not compared | 2026-11-12 |
| 99 | Qwen3.5-397B-A17BA | OVH AI Endpoints (GRA) | A | 3.1 s | 3.8 s | 29% | 2026-11-12 |
| 100 | GLM-5.1A | Z.ai (GLM / Zhipu) | A | 3.2 s | 3.7 s | 32% | 2026-11-12 |
| 101 | gpt-5.6-solB | OpenAI | B | 3.4 s | 5.5 s | 61% | 2026-11-12 |
| 102 | GLM-5A | Z.ai (GLM / Zhipu) | A | 3.5 s | 5.2 s | 53% | 2026-11-12 |
| 103 | gpt-6.1-solA | OpenAI | A | 4.1 s | 4.7 s | 65% | 2026-11-12 |
| 104 | GLM-4.7A | Z.ai (GLM / Zhipu) | A | 4.7 s | 6.8 s | 61% | 2026-11-12 |
| 105 | GLM-4.5A | Z.ai (GLM / Zhipu) | A | 6.8 s | 9.1 s | 56% | 2026-11-12 |
| 106 | GLM-4.6A | Z.ai (GLM / Zhipu) | A | 7.4 s | 24.8 s | 30% | 2026-11-12 |
| 107 | GLM-5 TurboB | Z.ai (GLM / Zhipu) | B | 7.6 s | 21.4 s | 34% | 2026-11-12 |
| 108 | Claude Sonnet 4.6A | Anthropic | A | 7.6 s | 7.7 s | 59% | 2026-11-12 |
| · | Deep Research Max Preview (Apr-21-2026)B | Google Gemini | B | Cannot be measured — answers on a different API | not measured | not compared | — |
| · | Deep Research Preview (Apr-21-2026)B | Google Gemini | B | Cannot be measured — answers on a different API | not measured | not compared | — |
| · | Deep Research Pro Preview (Dec-12-2025)B | Google Gemini | B | Cannot be measured — answers on a different API | not measured | not compared | — |
| · | Gemini 2.5 Computer Use Preview 10-2025B | Google Gemini | B | Cannot be measured — answers on a different API | not measured | not compared | — |
112 of 112 models · click a column to sort
“Cannot be measured” means the model does not answer on the chat API we time, or its provider uses a protocol our prober does not speak yet. Waiting will not fill these in.
Fast (under 500 ms)
Medium (500–1000 ms)
Slow (over 1000 ms)
Measured four times a day (02:00, 08:00, 14:00, 20:00 UTC) from Amsterdam.