Leaderboard: overall

Updated Jul. 25, 2025

  • Ranking: The position of the model in the leaderboard as ordered by Overall Score
  • Organization: The group responsible for the model or forecasts
  • Model: The LLM model & prompt info or the human group and forecast aggregation method
    • zero shot: used a zero-shot prompt
    • scratchpad: used a scratchpad prompt with instructions that outline a procedure the model should use to reason about the question
    • with freeze values: means that, for questions from market sources, the prompt was supplemented with the aggregate human forecast from the relevant platform on the day the question set was generated
    • with news: means that the prompt was supplemented with relevant news summaries obtained through an automated process
  • Dataset Score: The average Brier score across all questions sourced from datasets
  • Market Score (resolved): The average Brier score across all resolved questions sourced from prediction markets and forecast aggregation platforms
  • Market Score (unresolved): The average Brier score across all unresolved questions sourced from prediction markets and forecast aggregation platforms
  • Market Score (overall): The average Brier score across all questions sourced from prediction markets and forecast aggregation platforms
  • Overall Resolved Score: The average of the Dataset Score and the Market Score (resolved) columns
  • Overall Score: The average of the Dataset Score and the Market Score (overall) columns
  • Overall Score 95% CI: The 95% confidence interval for the Overall Score
  • Pairwise p-value comparing to No. 1 (bootstrapped): The p-value calculated by bootstrapping the differences in overall score between each model and the best forecaster (the group with rank 1) under the null hypothesis that there's no difference.
  • Pct. more accurate than No. 1: The percent of questions where this forecaster had a better overall score than the best forecaster (with rank 1)
  • Pct. imputed: The percent of questions for which this forecaster did not provide a forecast and hence had a forecast value imputed (0.5 for dataset questions and the aggregate human forecast on the forecast due date for questions sourced from prediction markets or forecast aggregation platforms)
Ranking Organization Model Dataset Score (N=6,562) Market Score (resolved) (N=536) Market Score (unresolved) (N=351) Market Score (overall) (N=887) Overall Resolved Score (N=7,098) Overall Score (N=7,449) Overall Score 95% CI Pairwise p-value comparing to No. 1 (bootstrapped) Pct. more accurate than No. 1 Pct. Imputed
1 Anthropic Claude-3-5-Sonnet-20240620 (scratchpad with freeze values) 0.176 0.113 0.033 0.082 0.145 0.129 [0.123, 0.135] 0% 0%
2 OpenAI GPT-4-Turbo-2024-04-09 (scratchpad with freeze values) 0.176 0.111 0.040 0.083 0.144 0.130 [0.124, 0.136] 0.368 39% 0%
3 OpenAI GPT-4o (scratchpad with freeze values) 0.191 0.098 0.034 0.073 0.145 0.132 [0.126, 0.137] 0.092 39% 0%
4 Google Gemini-1.5-Pro (scratchpad with freeze values) 0.166 0.136 0.067 0.109 0.151 0.137 [0.131, 0.143] <0.01 41% 0%
5 OpenAI GPT-4o (scratchpad with news with freeze values) 0.196 0.117 0.041 0.087 0.156 0.141 [0.135, 0.147] <0.001 36% 0%
6 Google Gemini-1.5-Pro (scratchpad with news with freeze values) 0.171 0.140 0.076 0.115 0.155 0.143 [0.137, 0.149] <0.001 40% 0%
7 Anthropic Claude-3-Opus-20240229 (zero shot with freeze values) 0.192 0.130 0.045 0.096 0.161 0.144 [0.138, 0.151] <0.001 41% 0%
8 Qwen Qwen1.5-110B-Chat (scratchpad with freeze values) 0.181 0.143 0.065 0.112 0.162 0.147 [0.141, 0.152] <0.001 35% 0%
9 OpenAI GPT-4-Turbo-2024-04-09 (scratchpad) 0.176 0.145 0.076 0.118 0.161 0.147 [0.142, 0.153] <0.001 36% 0%
10 OpenAI GPT-4-Turbo-2024-04-09 (zero shot with freeze values) 0.208 0.114 0.045 0.087 0.161 0.147 [0.141, 0.154] <0.001 39% 0%
11 Anthropic Claude-3-5-Sonnet-20240620 (zero shot with freeze values) 0.197 0.129 0.052 0.098 0.163 0.148 [0.141, 0.155] <0.001 40% 0%
12 Anthropic Claude-3-5-Sonnet-20240620 (scratchpad with news with freeze values) 0.191 0.135 0.060 0.105 0.163 0.148 [0.142, 0.154] <0.001 34% 0%
13 Google Gemini-1.5-Pro (scratchpad) 0.166 0.163 0.084 0.131 0.164 0.149 [0.143, 0.154] <0.001 40% 0%
14 Anthropic Claude-3-5-Sonnet-20240620 (scratchpad) 0.176 0.153 0.079 0.123 0.164 0.150 [0.144, 0.155] <0.001 3% 0%
15 ForecastBench Imputed Forecaster 0.250 0.068 0.024 0.051 0.159 0.150 [0.146, 0.155] <0.001 36% 100%
16 OpenAI GPT-4 (scratchpad with freeze values) 0.198 0.136 0.055 0.104 0.167 0.151 [0.145, 0.157] <0.001 34% 0%
17 Google Gemini-1.5-Pro (scratchpad with news) 0.171 0.158 0.091 0.131 0.164 0.151 [0.145, 0.157] <0.001 39% 0%
18 OpenAI GPT-4o (scratchpad) 0.191 0.149 0.066 0.116 0.170 0.153 [0.148, 0.159] <0.001 36% 0%
19 Google Gemini-1.5-Pro (zero shot with freeze values) 0.221 0.106 0.057 0.087 0.163 0.154 [0.148, 0.16] <0.001 34% 5%
20 OpenAI GPT-4-Turbo-2024-04-09 (scratchpad with news with freeze values) 0.216 0.118 0.058 0.094 0.167 0.155 [0.149, 0.161] <0.001 33% 0%
21 OpenAI GPT-4o (scratchpad with news) 0.196 0.145 0.074 0.117 0.170 0.156 [0.15, 0.162] <0.001 34% 0%
22 ForecastBench LLM Crowd (gpt-4o, claude-3.5-sonnet, gemini-1.5-pro) with news 0.244 0.093 0.038 0.071 0.168 0.157 [0.153, 0.162] <0.001 34% 88%
23 OpenAI GPT-4 (zero shot with freeze values) 0.226 0.119 0.046 0.090 0.172 0.158 [0.151, 0.164] <0.001 34% 0%
24 Qwen Qwen1.5-110B-Chat (scratchpad) 0.181 0.170 0.089 0.138 0.175 0.159 [0.154, 0.164] <0.001 34% 0%
25 ForecastBench LLM Crowd (gpt-4o, claude-3.5-sonnet, gemini-1.5-pro) with news 0.244 0.098 0.039 0.074 0.171 0.159 [0.155, 0.164] <0.001 34% 88%
26 Anthropic Claude-3-5-Sonnet-20240620 (scratchpad with news) 0.191 0.158 0.083 0.128 0.174 0.159 [0.154, 0.165] <0.001 34% 0%
27 ForecastBench LLM Crowd (gpt-4o, claude-3.5-sonnet, gemini-1.5-pro) with news 0.245 0.098 0.038 0.074 0.171 0.160 [0.155, 0.164] <0.001 33% 88%
28 Anthropic Claude-3-Opus-20240229 (scratchpad with freeze values) 0.206 0.146 0.068 0.115 0.176 0.161 [0.155, 0.167] <0.001 31% 0%
29 Mistral AI Mistral-Large-Latest (scratchpad with freeze values) 0.205 0.149 0.069 0.118 0.177 0.161 [0.156, 0.167] <0.001 29% 0%
30 Google Gemini-1.5-Pro (zero shot) 0.221 0.111 0.086 0.101 0.166 0.161 [0.155, 0.168] <0.001 33% 5%
31 OpenAI GPT-4 (scratchpad) 0.198 0.157 0.074 0.124 0.178 0.161 [0.156, 0.166] <0.001 32% 0%
32 Google Gemini-1.5-Pro (superforecaster with news 3) 0.189 0.165 0.088 0.135 0.177 0.162 [0.156, 0.168] <0.001 38% 0%
33 OpenAI GPT-4-Turbo-2024-04-09 (superforecaster with news 3) 0.208 0.147 0.079 0.120 0.177 0.164 [0.158, 0.169] <0.001 31% 13%
34 Anthropic Claude-3-5-Sonnet-20240620 (superforecaster with news 3) 0.196 0.165 0.083 0.132 0.180 0.164 [0.158, 0.17] <0.001 33% 3%
35 OpenAI GPT-4-Turbo-2024-04-09 (zero shot) 0.208 0.146 0.082 0.121 0.177 0.164 [0.158, 0.171] <0.001 36% 0%
36 Meta Llama-3-70b-Chat-Hf (zero shot with freeze values) 0.226 0.137 0.054 0.104 0.182 0.165 [0.158, 0.172] <0.001 31% 0%
37 Google Gemini-1.5-Flash (scratchpad with freeze values) 0.199 0.163 0.083 0.131 0.181 0.165 [0.158, 0.172] <0.001 35% 0%
38 Anthropic Claude-3-Opus-20240229 (zero shot) 0.192 0.168 0.097 0.140 0.180 0.166 [0.159, 0.173] <0.001 39% 0%
39 OpenAI GPT-4-Turbo-2024-04-09 (scratchpad with news) 0.216 0.136 0.086 0.116 0.176 0.166 [0.161, 0.172] <0.001 31% 0%
40 Anthropic Claude-3-5-Sonnet-20240620 (zero shot) 0.197 0.168 0.094 0.138 0.182 0.168 [0.161, 0.175] <0.001 38% 0%
41 Qwen Qwen1.5-110B-Chat (scratchpad with news with freeze values) 0.210 0.161 0.074 0.126 0.185 0.168 [0.162, 0.174] <0.001 30% 0%
42 Mistral AI Mixtral-8x22B-Instruct-V0.1 (scratchpad with freeze values) 0.212 0.152 0.082 0.125 0.182 0.169 [0.162, 0.175] <0.001 32% 0%
43 Google Gemini-1.5-Flash (zero shot with freeze values) 0.238 0.131 0.059 0.103 0.185 0.170 [0.163, 0.178] <0.001 39% 1%
44 Mistral AI Mistral-Large-Latest (zero shot with freeze values) 0.216 0.161 0.070 0.125 0.188 0.171 [0.163, 0.178] <0.001 32% 0%
45 OpenAI GPT-4o (superforecaster with news 3) 0.214 0.162 0.076 0.128 0.188 0.171 [0.165, 0.177] <0.001 31% 11%
46 Meta Llama-3-70b-Chat-Hf (scratchpad with freeze values) 0.232 0.144 0.062 0.112 0.188 0.172 [0.166, 0.178] <0.001 26% 0%
47 Anthropic Claude-2.1 (scratchpad) 0.241 0.128 0.066 0.104 0.185 0.172 [0.167, 0.178] <0.001 34% 7%
48 OpenAI GPT-4o (zero shot with freeze values) 0.228 0.149 0.076 0.120 0.189 0.174 [0.166, 0.182] <0.001 35% 1%
49 Google Gemini-1.5-Flash (scratchpad) 0.199 0.182 0.100 0.150 0.191 0.174 [0.168, 0.181] <0.001 34% 0%
50 Anthropic Claude-3-Opus-20240229 (scratchpad) 0.206 0.180 0.091 0.145 0.193 0.176 [0.17, 0.181] <0.001 31% 0%
51 OpenAI GPT-4o (superforecaster with news 1) 0.221 0.151 0.099 0.130 0.186 0.176 [0.169, 0.183] <0.001 35% 0%
52 Qwen Qwen1.5-110B-Chat (zero shot with freeze values) 0.230 0.155 0.071 0.122 0.193 0.176 [0.169, 0.183] <0.001 32% 0%
53 Anthropic Claude-3-5-Sonnet-20240620 (superforecaster with news 1) 0.215 0.167 0.092 0.138 0.191 0.176 [0.17, 0.183] <0.001 33% 27%
54 Meta Llama-3-8b-Chat-Hf (zero shot with freeze values) 0.223 0.143 0.111 0.130 0.183 0.177 [0.169, 0.184] <0.001 46% 0%
55 OpenAI GPT-4-Turbo-2024-04-09 (superforecaster with news 1) 0.221 0.164 0.084 0.133 0.193 0.177 [0.17, 0.183] <0.001 32% 0%
56 OpenAI GPT-4o (scratchpad with SECOND news) 0.236 0.145 0.075 0.118 0.191 0.177 [0.171, 0.183] <0.001 31% 18%
57 OpenAI GPT-4o (zero shot) 0.228 0.166 0.065 0.126 0.197 0.177 [0.17, 0.184] <0.001 33% 1%
58 Anthropic Claude-2.1 (scratchpad with freeze values) 0.241 0.151 0.056 0.113 0.196 0.177 [0.171, 0.184] <0.001 33% 5%
59 Qwen Qwen1.5-110B-Chat (scratchpad with news) 0.210 0.185 0.091 0.148 0.197 0.179 [0.173, 0.184] <0.001 29% 0%
60 Mistral AI Mistral-Large-Latest (scratchpad) 0.205 0.191 0.096 0.153 0.198 0.179 [0.174, 0.185] <0.001 28% 0%
61 Meta Llama-3-70b-Chat-Hf (zero shot) 0.226 0.159 0.096 0.134 0.192 0.180 [0.174, 0.186] <0.001 29% 0%
62 Anthropic Claude-3-5-Sonnet-20240620 (scratchpad with SECOND news) 0.225 0.167 0.086 0.135 0.196 0.180 [0.174, 0.186] <0.001 31% 2%
63 Mistral AI Mixtral-8x22B-Instruct-V0.1 (scratchpad) 0.212 0.187 0.091 0.149 0.200 0.181 [0.175, 0.186] <0.001 31% 0%
64 Google Gemini-1.5-Flash (scratchpad with news with freeze values) 0.228 0.170 0.087 0.137 0.199 0.183 [0.176, 0.19] <0.001 32% 0%
65 Anthropic Claude-2.1 (scratchpad with news) 0.243 0.163 0.065 0.124 0.203 0.184 [0.177, 0.19] <0.001 33% 6%
66 Qwen Qwen1.5-110B-Chat (superforecaster with news 1) 0.216 0.190 0.105 0.156 0.203 0.186 [0.179, 0.193] <0.001 33% 25%
67 Mistral AI Mixtral-8x22B-Instruct-V0.1 (zero shot with freeze values) 0.219 0.192 0.096 0.154 0.205 0.186 [0.178, 0.195] <0.001 30% 0%
68 Anthropic Claude-2.1 (scratchpad with news with freeze values) 0.243 0.165 0.074 0.129 0.204 0.186 [0.18, 0.193] <0.001 32% 5%
69 Anthropic Claude-3-Opus-20240229 (superforecaster with news 1) 0.223 0.182 0.101 0.150 0.203 0.187 [0.18, 0.193] <0.001 30% 14%
70 Mistral AI Mixtral-8x22B-Instruct-V0.1 (zero shot) 0.219 0.184 0.109 0.154 0.202 0.187 [0.18, 0.194] <0.001 30% 0%
71 Qwen Qwen1.5-110B-Chat (superforecaster with news 3) 0.222 0.192 0.088 0.151 0.207 0.187 [0.181, 0.192] <0.001 29% 6%
72 Google Gemini-1.5-Pro (superforecaster with news 1) 0.222 0.184 0.102 0.152 0.203 0.187 [0.18, 0.194] <0.001 34% 7%
73 Anthropic Claude-3-Opus-20240229 (scratchpad with news with freeze values) 0.228 0.183 0.092 0.147 0.205 0.187 [0.181, 0.193] <0.001 27% 0%
74 OpenAI GPT-4 (zero shot) 0.226 0.192 0.085 0.149 0.209 0.187 [0.182, 0.193] <0.001 30% 0%
75 Qwen Qwen1.5-110B-Chat (zero shot) 0.230 0.180 0.103 0.150 0.205 0.190 [0.184, 0.196] <0.001 31% 0%
76 Meta Llama-3-8b-Chat-Hf (zero shot) 0.223 0.180 0.129 0.160 0.201 0.191 [0.183, 0.2] <0.001 46% 0%
77 Meta Llama-3-70b-Chat-Hf (scratchpad) 0.232 0.190 0.092 0.152 0.211 0.192 [0.187, 0.197] <0.001 26% 0%
78 Google Gemini-1.5-Flash (scratchpad with news) 0.228 0.187 0.108 0.156 0.208 0.192 [0.186, 0.198] <0.001 31% 0%
79 Mistral AI Mixtral-8x22B-Instruct-V0.1 (scratchpad with news with freeze values) 0.236 0.189 0.088 0.149 0.213 0.193 [0.187, 0.199] <0.001 29% 0%
80 Mistral AI Mistral-Large-Latest (zero shot) 0.216 0.199 0.126 0.170 0.207 0.193 [0.186, 0.2] <0.001 31% 0%
81 Anthropic Claude-3-Opus-20240229 (superforecaster with news 3) 0.222 0.202 0.105 0.164 0.212 0.193 [0.187, 0.199] <0.001 30% 9%
82 Mistral AI Mixtral-8x22B-Instruct-V0.1 (superforecaster with news 1) 0.237 0.172 0.114 0.149 0.205 0.193 [0.186, 0.2] <0.001 32% 32%
83 Google Gemini-1.5-Flash (superforecaster with news 2) 0.253 0.151 0.108 0.134 0.202 0.193 [0.187, 0.2] <0.001 29% 6%
84 Anthropic Claude-3-Opus-20240229 (scratchpad with news) 0.228 0.199 0.102 0.161 0.214 0.194 [0.189, 0.2] <0.001 27% 0%
85 Mistral AI Mixtral-8x22B-Instruct-V0.1 (superforecaster with news 3) 0.233 0.197 0.098 0.158 0.215 0.196 [0.19, 0.201] <0.001 30% 18%
86 Anthropic Claude-3-5-Sonnet-20240620 (superforecaster with news 2) 0.229 0.201 0.104 0.163 0.215 0.196 [0.189, 0.203] <0.001 30% 1%
87 Mistral AI Mixtral-8x22B-Instruct-V0.1 (scratchpad with news) 0.236 0.194 0.098 0.156 0.215 0.196 [0.191, 0.201] <0.001 28% 0%
88 Anthropic Claude-3-Opus-20240229 (superforecaster with news 2) 0.233 0.194 0.107 0.160 0.214 0.196 [0.19, 0.203] <0.001 28% 1%
89 Mistral AI Mixtral-8x7B-Instruct-V0.1 (zero shot with freeze values) 0.267 0.154 0.096 0.131 0.210 0.199 [0.19, 0.207] <0.001 38% 0%
90 Anthropic Claude-2.1 (superforecaster with news 2) 0.271 0.168 0.069 0.129 0.220 0.200 [0.193, 0.207] <0.001 34% 10%
91 Mistral AI Mixtral-8x7B-Instruct-V0.1 (scratchpad with freeze values) 0.254 0.182 0.093 0.147 0.218 0.201 [0.193, 0.208] <0.001 28% 3%
92 Mistral AI Mixtral-8x7B-Instruct-V0.1 (scratchpad) 0.254 0.179 0.098 0.147 0.217 0.201 [0.194, 0.207] <0.001 28% 4%
93 Google Gemini-1.5-Flash (superforecaster with news 3) 0.234 0.208 0.105 0.167 0.221 0.201 [0.194, 0.207] <0.001 29% 15%
94 Google Gemini-1.5-Pro (superforecaster with news 2) 0.244 0.194 0.104 0.158 0.219 0.201 [0.193, 0.209] <0.001 33% 1%
95 Anthropic Claude-2.1 (zero shot with freeze values) 0.247 0.200 0.101 0.161 0.223 0.204 [0.195, 0.212] <0.001 32% 0%
96 OpenAI GPT-4o (superforecaster with news 2) 0.262 0.188 0.090 0.149 0.225 0.206 [0.199, 0.213] <0.001 31% 6%
97 OpenAI GPT-4-Turbo-2024-04-09 (superforecaster with news 2) 0.250 0.199 0.108 0.163 0.225 0.207 [0.199, 0.214] <0.001 31% 11%
98 Anthropic Claude-2.1 (superforecaster with news 1) 0.269 0.183 0.085 0.145 0.226 0.207 [0.2, 0.214] <0.001 30% 16%
99 Qwen Qwen1.5-110B-Chat (superforecaster with news 2) 0.245 0.213 0.104 0.170 0.229 0.207 [0.201, 0.214] <0.001 27% 6%
100 Google Gemini-1.5-Flash (zero shot) 0.238 0.197 0.147 0.177 0.218 0.208 [0.2, 0.216] <0.001 36% 1%
101 Mistral AI Mixtral-8x7B-Instruct-V0.1 (superforecaster with news 3) 0.280 0.160 0.099 0.136 0.220 0.208 [0.202, 0.214] <0.001 28% 7%
102 Mistral AI Mixtral-8x22B-Instruct-V0.1 (superforecaster with news 2) 0.251 0.209 0.104 0.167 0.230 0.209 [0.203, 0.215] <0.001 27% 3%
103 Mistral AI Mistral-Large-Latest (scratchpad with news with freeze values) 0.258 0.212 0.099 0.167 0.235 0.213 [0.206, 0.219] <0.001 24% 0%
104 Mistral AI Mistral-Large-Latest (superforecaster with news 1) 0.251 0.213 0.122 0.177 0.232 0.214 [0.207, 0.222] <0.001 30% 41%
105 Mistral AI Mistral-Large-Latest (scratchpad with news) 0.258 0.217 0.112 0.176 0.238 0.217 [0.211, 0.223] <0.001 24% 0%
106 Mistral AI Mistral-Large-Latest (superforecaster with news 2) 0.243 0.237 0.122 0.192 0.240 0.217 [0.21, 0.225] <0.001 30% 16%
107 Anthropic Claude-2.1 (superforecaster with news 3) 0.267 0.216 0.106 0.173 0.242 0.220 [0.213, 0.226] <0.001 28% 8%
108 Anthropic Claude-2.1 (zero shot) 0.247 0.236 0.130 0.194 0.241 0.220 [0.214, 0.227] <0.001 30% 0%
109 Mistral AI Mixtral-8x7B-Instruct-V0.1 (zero shot) 0.267 0.206 0.128 0.175 0.236 0.221 [0.212, 0.23] <0.001 37% 0%
110 Mistral AI Mixtral-8x7B-Instruct-V0.1 (scratchpad with news) 0.291 0.181 0.104 0.151 0.236 0.221 [0.215, 0.227] <0.001 26% 5%
111 Mistral AI Mixtral-8x7B-Instruct-V0.1 (scratchpad with news with freeze values) 0.291 0.179 0.111 0.152 0.235 0.222 [0.215, 0.228] <0.001 26% 5%
112 Mistral AI Mixtral-8x7B-Instruct-V0.1 (superforecaster with news 1) 0.303 0.168 0.099 0.141 0.236 0.222 [0.215, 0.229] <0.001 28% 6%
113 ForecastBench Always 0.5 0.250 0.250 0.113 0.196 0.250 0.223 [0.22, 0.226] <0.001 30% 0%
114 Google Gemini-1.5-Flash (superforecaster with news 1) 0.248 0.236 0.145 0.200 0.242 0.224 [0.216, 0.232] <0.001 29% 28%
115 Anthropic Claude-3-Haiku-20240307 (zero shot with freeze values) 0.281 0.212 0.102 0.168 0.247 0.225 [0.218, 0.232] <0.001 29% 1%
116 Meta Llama-3-8b-Chat-Hf (scratchpad with freeze values) 0.277 0.213 0.116 0.175 0.245 0.226 [0.219, 0.232] <0.001 25% 0%
117 Meta Llama-2-70b-Chat-Hf (scratchpad with freeze values) 0.269 0.230 0.116 0.185 0.249 0.227 [0.221, 0.232] <0.001 26% 0%
118 Mistral AI Mistral-Large-Latest (superforecaster with news 3) 0.272 0.230 0.108 0.182 0.251 0.227 [0.22, 0.233] <0.001 25% 8%
119 Meta Llama-2-70b-Chat-Hf (scratchpad) 0.269 0.240 0.108 0.188 0.254 0.228 [0.223, 0.234] <0.001 26% 0%
120 Mistral AI Mixtral-8x7B-Instruct-V0.1 (superforecaster with news 2) 0.299 0.188 0.122 0.162 0.243 0.230 [0.223, 0.237] <0.001 28% 10%
121 Anthropic Claude-3-Haiku-20240307 (scratchpad with freeze values) 0.280 0.233 0.114 0.186 0.256 0.233 [0.227, 0.239] <0.001 23% 0%
122 ForecastBench Always 0 0.268 0.203 0.201 0.202 0.236 0.235 [0.222, 0.248] <0.001 71% 0%
123 Anthropic Claude-3-Haiku-20240307 (zero shot) 0.281 0.240 0.121 0.193 0.261 0.237 [0.231, 0.243] <0.001 29% 1%
124 Anthropic Claude-3-Haiku-20240307 (superforecaster with news 2) 0.271 0.265 0.118 0.207 0.268 0.239 [0.232, 0.245] <0.001 25% 9%
125 Meta Llama-3-8b-Chat-Hf (scratchpad) 0.277 0.248 0.131 0.202 0.262 0.239 [0.233, 0.245] <0.001 25% 0%
126 Anthropic Claude-3-Haiku-20240307 (scratchpad) 0.280 0.263 0.129 0.210 0.272 0.245 [0.24, 0.25] <0.001 22% 0%
127 Anthropic Claude-3-Haiku-20240307 (scratchpad with news with freeze values) 0.309 0.245 0.117 0.194 0.277 0.252 [0.246, 0.257] <0.001 22% 0%
128 Anthropic Claude-3-Haiku-20240307 (superforecaster with news 3) 0.288 0.281 0.135 0.223 0.285 0.256 [0.25, 0.261] <0.001 24% 24%
129 Anthropic Claude-3-Haiku-20240307 (scratchpad with news) 0.309 0.266 0.124 0.210 0.287 0.259 [0.254, 0.265] <0.001 22% 0%
130 OpenAI GPT-3.5-Turbo-0125 (scratchpad with freeze values) 0.300 0.276 0.139 0.222 0.288 0.261 [0.255, 0.267] <0.001 23% 0%
131 OpenAI GPT-3.5-Turbo-0125 (scratchpad) 0.300 0.282 0.139 0.225 0.291 0.263 [0.257, 0.269] <0.001 23% 0%
132 Meta Llama-2-70b-Chat-Hf (zero shot with freeze values) 0.309 0.289 0.153 0.235 0.299 0.272 [0.263, 0.282] <0.001 29% 0%
133 Anthropic Claude-3-Haiku-20240307 (superforecaster with news 1) 0.302 0.324 0.150 0.255 0.313 0.279 [0.271, 0.286] <0.001 24% 10%
134 Meta Llama-2-70b-Chat-Hf (zero shot) 0.309 0.324 0.159 0.258 0.317 0.284 [0.276, 0.291] <0.001 29% 0%
135 ForecastBench Random Uniform 0.338 0.326 0.188 0.272 0.332 0.305 [0.295, 0.315] <0.001 32% 0%
136 OpenAI GPT-3.5-Turbo-0125 (zero shot with freeze values) 0.440 0.247 0.144 0.206 0.343 0.323 [0.313, 0.333] <0.001 26% 0%
137 OpenAI GPT-3.5-Turbo-0125 (zero shot) 0.440 0.337 0.220 0.291 0.388 0.365 [0.356, 0.374] <0.001 24% 0%
138 ForecastBench Always 1 0.732 0.797 0.525 0.689 0.764 0.711 [0.696, 0.725] <0.001 26% 0%