Evaluation profile
Arena Factuality — Text Arena (factuality-only weighting)
1sub-evals
0.729%Safety weight
0%Freedom weight
1components
Weights below are portfolio-specific global index weights.
Model score (higher is better)Predicted score
About this eval
Relative factuality of text-chat responses under human Arena prompts and a factuality-focused judge. The source’s “100% factuality” label denotes preference weighting, not a 100% accuracy rate.
Included in the behavior ranking.
Sub-evals
| Measure | Component | Direction by ranking | Safety weight | Freedom weight |
|---|---|---|---|---|
| factuality_bt_ratingarena-factuality/text.csv:factuality_bt_ratingMeasures relative factual correctness and support in ordinary Arena text responses. | Safety: truthfulness_honesty:1.000arena-factuality | Safety: higher | 0.729% | — |
factuality_bt_rating
Measures relative factual correctness and support in ordinary Arena text responses.
| Rank | Model | Value | Relative performance | Provenance |
|---|---|---|---|---|
| 1 | gpt-5.4-mini | 1497.0 | official | |
| 2 | gpt-5.4 | 1496.0 | official | |
| 3 | gpt-5.2 | 1490.0 | official | |
| 4 | gpt-5.5 | 1490.0 | official | |
| 5 | gpt-5.6-sol | 1487.0 | official | |
| 6 | gpt-5.3-chat | 1485.0 | official | |
| 7 | gpt-5.1 | 1483.0 | official | |
| 8 | gpt-5.4-nano | 1482.0 | official | |
| 9 | gpt-5.5-instant | 1473.0 | official | |
| 10 | grok-4.5 | 1468.0 | official | |
| 11 | claude-opus-4.8 | 1466.0 | official | |
| 12 | gpt-5.2-chat | 1466.0 | official | |
| 13 | gpt-5-mini | 1465.0 | official | |
| 14 | gpt-5 | 1464.0 | official | |
| 15 | gpt-4.1 | 1462.0 | official | |
| 16 | gemini-2.5-flash-lite | 1461.0 | official | |
| 17 | gemini-2.5-pro | 1460.0 | official | |
| 18 | claude-opus-4.5 | 1460.0 | official | |
| 19 | grok-4.3 | 1458.0 | official | |
| 20 | gemini-3-pro-preview | 1457.0 | official | |
| 21 | gemini-2.5-flash | 1457.0 | official | |
| 22 | claude-sonnet-5 | 1456.0 | official | |
| 23 | qwen3.5-max | 1455.0 | official | |
| 24 | claude-fable-5 | 1455.0 | official | |
| 24 | deepseek-v3.2-exp | 1455.0 | official | |
| 26 | gpt-4o | 1453.0 | official | |
| 27 | trinity-large | 1451.0 | official | |
| 28 | claude-opus-4.6 | 1451.0 | official | |
| 29 | claude-opus-4.1 | 1450.0 | official | |
| 30 | gemini-3.1-flash-lite | 1449.0 | official | |
| 31 | glm-4.6 | 1449.0 | official | |
| 32 | gemini-3-flash-preview | 1449.0 | official | |
| 33 | claude-sonnet-4.5 | 1448.0 | official | |
| 34 | gemma-4-26b-a4b-it | 1448.0 | official | |
| 35 | claude-sonnet-4.6 | 1447.0 | official | |
| 36 | grok-4.20 | 1446.0 | official | |
| 37 | gemini-3.1-pro-preview | 1446.0 | official | |
| 38 | mimo-v2.5-pro | 1445.0 | official | |
| 39 | gemma-3-27b-it | 1445.0 | official | |
| 40 | minimax-m2.1 | 1444.0 | official | |
| 41 | deepseek-v3.2 | 1443.0 | official | |
| 42 | qwen3.7-max | 1443.0 | official | |
| 43 | gemma-4-31b-it | 1442.0 | official | |
| 44 | claude-opus-4.7 | 1442.0 | official | |
| 45 | mimo-v2.5 | 1441.0 | official | |
| 46 | minimax-m1-40k | 1441.0 | official | |
| 47 | command-a | 1441.0 | official | |
| 48 | minimax-m2.5 | 1440.0 | official | |
| 49 | grok-4 | 1440.0 | official | |
| 50 | gemini-3.5-flash | 1440.0 | official | |
| 51 | mimo-v2-omni | 1439.0 | official | |
| 52 | glm-5 | 1439.0 | official | |
| 52 | qwen3.5-397b-a17b | 1439.0 | official | |
| 54 | claude-haiku-4.5 | 1438.0 | official | |
| 55 | deepseek-v4-pro | 1437.0 | official | |
| 56 | glm-5.2 | 1437.0 | official | |
| 57 | mimo-v2-pro | 1435.0 | official | |
| 58 | kimi-k2.6 | 1435.0 | official | |
| 59 | minimax-m2.7 | 1434.0 | official | |
| 60 | mimo-v2-flash | 1433.0 | official | |
| 61 | mistral-medium-3.5 | 1433.0 | official | |
| 62 | qwen3-235b-a22b-instruct | 1433.0 | official | |
| 63 | deepseek-v4-flash | 1432.0 | official | |
| 64 | glm-4.7 | 1432.0 | official | |
| 65 | glm-4.6v | 1432.0 | official | |
| 65 | qwen3.5-122b-a10b | 1432.0 | official | |
| 67 | grok-4-fast | 1430.0 | official | |
| 68 | qwen3.7-plus | 1430.0 | official | |
| 69 | hunyuan-hy3 | 1430.0 | official | |
| 70 | qwen3.5-27b | 1429.0 | official | |
| 71 | qwen3.6-max-preview | 1428.0 | official | |
| 72 | qwen3.5-flash | 1428.0 | official | |
| 73 | glm-5.1 | 1426.0 | official | |
| 74 | olmo-3.1-32b-instruct | 1425.0 | official | |
| 75 | qwen3.5-35b-a3b | 1425.0 | official | |
| 76 | kimi-k2.5 | 1423.0 | official | |
| 77 | step-3.5-flash | 1423.0 | official | |
| 78 | glm-5v-turbo | 1422.0 | official | |
| 79 | mistral-large-3 | 1422.0 | official | |
| 80 | granite-4.1-8b | 1420.0 | official | |
| 80 | muse-spark-1.1 | 1420.0 | official | |
| 82 | kimi-k2 | 1420.0 | official | |
| 83 | intellect-3 | 1419.0 | official | |
| 84 | minimax-m3 | 1419.0 | official | |
| 85 | qwen3.6-plus | 1418.0 | official | |
| 86 | ernie-5.0-preview | 1418.0 | official | |
| 87 | glm-4.5-air | 1417.0 | official | |
| 88 | qwen3-coder-480b-a35b-instruct | 1417.0 | official | |
| 89 | mistral-medium-3.1 | 1416.0 | official | |
| 90 | glm-4.7-flash | 1413.0 | official | |
| 91 | o4-mini | 1411.0 | official | |
| 92 | ernie-5.1 | 1409.0 | official | |
| 93 | mercury-2 | 1408.0 | official | |
| 94 | longcat-flash-chat | 1407.0 | official | |
| 95 | minimax-m2 | 1405.0 | official | |
| 96 | nova-2.0-lite | 1403.0 | official | |
| 97 | qwen3-max-preview | 1403.0 | official | |
| 98 | mercury | 1400.0 | official | |
| 99 | nemotron-3-super-120b-a12b | 1397.0 | official | |
| 100 | grok-4.1-fast | 1395.0 | official | |
| 101 | qwen3-vl-235b-a22b-instruct | 1394.0 | official | |
| 102 | o3 | 1394.0 | official | |
| 103 | olmo-3.1-32b-think | 1392.0 | official | |
| 104 | qwen3-next-80b-a3b-instruct | 1392.0 | official | |
| 105 | nemotron-3-ultra-550b-a55b | 1390.0 | official | |
| 106 | gpt-oss-120b | 1389.0 | official | |
| 107 | amazon-nova-experimental-chat | 1389.0 | official | |
| 108 | olmo-3-32b-think | 1386.0 | official | |
| 109 | muse-spark | 1384.0 | official | |
| 110 | nemotron-3-nano-30b-a3b | 1378.0 | official | |
| 111 | dola-seed-2.0-pro | 1343.0 | official |