Memory Recursive Summarization
BFCL · 50 models
| Rank | Model | Score |
|---|---|---|
| 1 | Claude-Sonnet-4-5-20250929 (FC) | 83.23 |
| 2 | Claude-Opus-4-5-20251101 (FC) | 77.42 |
| 3 | o3-2025-04-16 (FC) | 72.9 |
| 4 | o3-2025-04-16 (Prompt) | 71.61 |
| 5 | GLM-4.6 (FC thinking) | 67.1 |
| 6 | Gemini-3-Pro-Preview (Prompt) | 63.23 |
| 6 | Grok-4-1-fast-reasoning (FC) | 63.23 |
| 8 | GPT-5.2-2025-12-11 (FC) | 60.65 |
| 9 | DeepSeek-V3.2-Exp (FC) | 59.35 |
| 10 | o4-mini-2025-04-16 (FC) | 58.71 |
| 11 | o4-mini-2025-04-16 (Prompt) | 58.06 |
| 12 | Claude-Haiku-4-5-20251001 (FC) | 56.13 |
| 13 | Gemini-2.5-Flash (Prompt) | 55.48 |
| 14 | Gemini-2.5-Flash (FC) | 54.19 |
| 15 | GPT-5-mini-2025-08-07 (FC) | 52.26 |
| 16 | Gemini-2.5-Flash-Lite (FC) | 51.61 |
| 16 | Grok-4-0709 (FC) | 51.61 |
| 18 | Falcon3-10B-Instruct (FC) | 50.97 |
| 18 | Gemini-3-Pro-Preview (FC) | 50.97 |
| 20 | Gemma-3-12b-it (Prompt) | 49.03 |
| 21 | Grok-4-0709 (Prompt) | 48.39 |
| 22 | Command A Reasoning (FC) | 46.45 |
| 23 | Qwen3-235B-A22B-Instruct-2507 (FC) | 45.81 |
| 23 | xLAM-2-32b-fc-r (FC) | 45.81 |
| 25 | Moonshotai-Kimi-K2-Instruct (FC) | 45.16 |
| 26 | Nanbeige4-3B-Thinking-2511 (FC) | 44.52 |
| 26 | Qwen3-4B-Instruct-2507 (Prompt) | 44.52 |
| 28 | GPT-4.1-mini-2025-04-14 (FC) | 41.94 |
| 28 | Qwen3-32B (FC) | 41.94 |
| 30 | Command A (FC) | 40 |
| 31 | DeepSeek-V3.2-Exp (Prompt + Thinking) | 39.35 |
| 31 | Nanbeige3.5-Pro-Thinking (FC) | 39.35 |
| 33 | Falcon3-7B-Instruct (FC) | 38.71 |
| 33 | GPT-4.1-mini-2025-04-14 (Prompt) | 38.71 |
| 33 | GPT-5-mini-2025-08-07 (Prompt) | 38.71 |
| 36 | Grok-4-1-fast-non-reasoning (FC) | 38.06 |
| 37 | GPT-4.1-2025-04-14 (FC) | 37.42 |
| 38 | GPT-4.1-2025-04-14 (Prompt) | 35.48 |
| 38 | Gemma-3-27b-it (Prompt) | 35.48 |
| 40 | Qwen3-14B (FC) | 34.84 |
| 40 | Qwen3-30B-A3B-Instruct-2507 (FC) | 34.84 |
| 42 | Mistral-Medium-2505 | 34.19 |
| 43 | Mistral-Medium-2505 (FC) | 33.55 |
| 43 | Qwen3-235B-A22B-Instruct-2507 (Prompt) | 33.55 |
| 43 | ToolACE-2-8B (FC) | 33.55 |
| 46 | Mistral-small-2506 (FC) | 31.61 |
| 46 | Phi-4 (Prompt) | 31.61 |
| 46 | Qwen3-8B (FC) | 31.61 |
| 49 | Mistral-Small-2506 (Prompt) | 30.97 |
| 50 | Gemini-2.5-Flash-Lite (Prompt) | 29.68 |