AI BENCHY Failures
Wrong answer Failures
See which AI models run into Wrong answer most often, so you can spot reliability risks before choosing one. Sort by: Response Time (avg) ↑.
| Rank | Model | Company | Wrong answer Count | Avg Score | Tests Correct | Response Time (avg) |
|---|---|---|---|---|---|---|
| #51 | Mercury 2 none | Inception | 11 | 3.4 | 4/16 | 596ms |
| #55 | LFM2-24B-A2B none | Liquid | 9 | 2.6 | 1/16 | 811ms |
| #38 | Gemini 2.5 Flash none | 9 | 5.2 | 6/16 | 923ms | |
| #22 | Gemini 3.1 Flash Lite Preview none | 4 | 7.1 | 10/16 | 1.33s | |
| #44 | GPT-5.4 none | OpenAI | 9 | 4.5 | 6/16 | 1.48s |
| #20 | Gemini 3 Flash Preview none | 5 | 7.2 | 11/16 | 1.75s | |
| #41 | Qwen3.5-27B none | Qwen | 9 | 4.9 | 5/16 | 1.75s |
| #53 | Grok 4.1 Fast none | X AI | 11 | 2.9 | 3/16 | 1.90s |
| #47 | GPT-4o-mini none | OpenAI | 11 | 4.0 | 4/16 | 2.07s |
| #36 | Mercury 2 medium | Inception | 5 | 5.3 | 7/16 | 2.36s |
| #29 | Qwen3.5 Plus 2026-02-15 none | Qwen | 7 | 6.2 | 9/16 | 2.65s |
| #54 | MiMo-V2-Flash none | Xiaomi | 10 | 2.9 | 3/16 | 2.97s |
| #49 | GLM 4.7 Flash none | Z.ai | 9 | 3.9 | 4/16 | 2.99s |
| #45 | Trinity Large Preview none | Arcee AI | 9 | 4.2 | 5/16 | 3.15s |
| #17 | Gemini 3.1 Flash Lite Preview low | 4 | 7.3 | 11/16 | 3.36s | |
| #37 | Qwen3.5-Flash none | Qwen | 8 | 5.2 | 7/16 | 3.54s |
| #40 | Qwen3.5-122B-A10B none | Qwen | 9 | 5.0 | 6/16 | 3.72s |
| #12 | Gemini 3.1 Flash Lite Preview medium | 4 | 7.5 | 11/16 | 3.83s | |
| #31 | GLM 5 none | Z.ai | 7 | 6.0 | 9/16 | 4.03s |
| #42 | Qwen3.5-35B-A3B none | Qwen | 8 | 4.7 | 6/16 | 4.10s |
| #25 | Claude Sonnet 4.6 none | Anthropic | 2 | 6.8 | 10/16 | 5.57s |
| #19 | GPT-5.3 Chat none | OpenAI | 4 | 7.3 | 10/16 | 5.96s |
| #5 | Gemini 3 Flash Preview low | 3 | 8.2 | 13/16 | 6.11s | |
| #15 | GPT-5.2 Chat none | OpenAI | 4 | 7.4 | 11/16 | 7.03s |
| #6 | Gemini 3 Pro Preview medium | 3 | 8.2 | 13/16 | 7.15s | |
| #11 | Claude Sonnet 4.6 medium | Anthropic | 1 | 7.7 | 12/16 | 11.2s |
| #48 | Qwen3 Coder Next none | Qwen | 10 | 4.0 | 4/16 | 11.7s |
| #46 | Kimi K2.5 none | Moonshot AI | 11 | 4.1 | 5/16 | 11.9s |
| #16 | Gemini 2.5 Flash medium | 4 | 7.4 | 11/16 | 12.4s | |
| #50 | Qwen3 Coder Next medium | Qwen | 8 | 3.5 | 3/16 | 12.5s |
| #33 | DeepSeek V3.2 none | DeepSeek | 6 | 5.5 | 7/16 | 12.9s |
| #27 | GPT-5.2 medium | OpenAI | 1 | 6.5 | 10/16 | 15.3s |
| #14 | GLM 5 medium | Z.ai | 2 | 7.4 | 11/16 | 16.2s |
| #3 | GPT-5.3-Codex medium | OpenAI | 2 | 8.4 | 12/16 | 16.6s |
| #2 | Gemini 3.1 Pro Preview medium | 1 | 9.4 | 15/16 | 16.6s | |
| #39 | gpt-oss-120b medium | OpenAI | 5 | 5.1 | 7/16 | 16.7s |
| #9 | GPT-5.4 medium | OpenAI | 2 | 8.0 | 12/16 | 20.1s |
| #26 | Claude Opus 4.6 medium | Anthropic | 2 | 6.6 | 10/16 | 22.9s |
| #32 | GPT-5 Mini medium | OpenAI | 3 | 6.0 | 8/16 | 25.1s |
| #21 | MiMo-V2-Flash medium | Xiaomi | 3 | 7.2 | 11/16 | 25.3s |
| #30 | Grok 4.1 Fast medium | X AI | 2 | 6.2 | 9/16 | 26.3s |
| #13 | Step 3.5 Flash medium | Stepfun | 3 | 7.4 | 10/16 | 29.1s |
| #10 | Qwen3.5-122B-A10B medium | Qwen | 3 | 7.7 | 12/16 | 29.7s |
| #4 | Qwen3.5 Plus 2026-02-15 medium | Qwen | 1 | 8.3 | 13/16 | 34.5s |
| #52 | GLM 4.7 Flash medium | Z.ai | 7 | 3.1 | 4/16 | 36.8s |
| #18 | DeepSeek V3.2 medium | DeepSeek | 3 | 7.3 | 11/16 | 39.5s |
| #43 | MiniMax M2.5 medium | Minimax | 5 | 4.7 | 5/16 | 43.0s |
| #35 | Qwen3.5-35B-A3B medium | Qwen | 2 | 5.5 | 8/16 | 43.9s |
| #34 | GPT-5 Nano medium | OpenAI | 5 | 5.5 | 7/16 | 47.9s |
| #7 | Qwen3.5-27B medium | Qwen | 1 | 8.2 | 12/16 | 52.1s |
| #23 | Seed-2.0-Mini medium | Bytedance Seed | 1 | 6.9 | 10/16 | 65.1s |
| #8 | Gemini 3.1 Flash Lite Preview high | 3 | 8.2 | 12/16 | 68.8s | |
| #28 | Kimi K2.5 medium | Moonshot AI | 3 | 6.4 | 9/16 | 69.8s |
| #24 | Qwen3.5-Flash medium | Qwen | 1 | 6.9 | 10/16 | 70.8s |