Leaderboard: overall

Updated Jan. 18, 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=4,792) Market Score (resolved) (N=502) Market Score (unresolved) (N=390) Market Score (overall) (N=892) Overall Resolved Score (N=5,294) Overall Score (N=5,684) 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) 0.175 0.111 0.034 0.077 0.143 0.126 [0.12, 0.132] 0% 0%
2 OpenAI GPT-4-Turbo-2024-04-09 (scratchpad with freeze values) 0.175 0.105 0.044 0.079 0.140 0.127 [0.121, 0.133] 0.326 44% 0%
3 OpenAI GPT-4o (scratchpad with freeze values) 0.193 0.095 0.035 0.068 0.144 0.131 [0.125, 0.136] <0.01 44% 0%
4 Google Gemini-1.5-Pro (scratchpad with freeze values) 0.165 0.134 0.065 0.104 0.149 0.134 [0.128, 0.14] <0.001 35% 0%
5 OpenAI GPT-4o (scratchpad with news with freeze values) 0.194 0.112 0.045 0.083 0.153 0.138 [0.132, 0.144] <0.001 40% 0%
6 Google Gemini-1.5-Pro (scratchpad with news with freeze values) 0.169 0.138 0.074 0.110 0.153 0.140 [0.134, 0.145] <0.001 34% 1%
7 Anthropic Claude-3-Opus-20240229 (zero shot with freeze values) 0.189 0.126 0.048 0.092 0.157 0.140 [0.133, 0.147] <0.001 41% 0%
8 OpenAI GPT-4-Turbo-2024-04-09 (zero shot with freeze values) 0.203 0.110 0.047 0.082 0.156 0.143 [0.136, 0.149] <0.001 42% 0%
9 Qwen Qwen1.5-110B-Chat (scratchpad with freeze values) 0.179 0.139 0.067 0.107 0.159 0.143 [0.137, 0.149] <0.001 31% 1%
10 OpenAI GPT-4-Turbo-2024-04-09 (scratchpad) 0.175 0.144 0.075 0.114 0.159 0.144 [0.139, 0.15] <0.001 31% 0%
11 Anthropic Claude-3-5-Sonnet-20240620 (scratchpad with news with freeze values) 0.189 0.132 0.059 0.100 0.160 0.144 [0.138, 0.151] <0.001 33% 0%
12 Google Gemini-1.5-Pro (scratchpad) 0.165 0.162 0.082 0.127 0.163 0.146 [0.14, 0.152] <0.001 32% 0%
13 Anthropic Claude-3-5-Sonnet-20240620 (zero shot with freeze values) 0.200 0.123 0.055 0.093 0.161 0.146 [0.139, 0.153] <0.001 42% 0%
14 Anthropic Claude-3-5-Sonnet-20240620 (scratchpad) 0.175 0.151 0.078 0.119 0.163 0.147 [0.141, 0.153] <0.001 10% 0%
15 OpenAI GPT-4 (scratchpad with freeze values) 0.197 0.133 0.055 0.099 0.165 0.148 [0.141, 0.154] <0.001 37% 0%
16 Google Gemini-1.5-Pro (scratchpad with news) 0.169 0.156 0.090 0.127 0.163 0.148 [0.142, 0.154] <0.001 32% 1%
17 ForecastBench Imputed Forecaster 0.250 0.063 0.025 0.047 0.157 0.148 [0.145, 0.152] <0.001 48% 100%
18 OpenAI GPT-4-Turbo-2024-04-09 (scratchpad with news with freeze values) 0.211 0.112 0.059 0.089 0.161 0.150 [0.144, 0.156] <0.001 36% 0%
19 Google Gemini-1.5-Pro (zero shot with freeze values) 0.219 0.102 0.056 0.082 0.161 0.151 [0.144, 0.157] <0.001 40% 19%
20 OpenAI GPT-4o (scratchpad) 0.193 0.145 0.071 0.113 0.169 0.153 [0.147, 0.159] <0.001 31% 1%
21 OpenAI GPT-4o (scratchpad with news) 0.194 0.139 0.078 0.112 0.166 0.153 [0.147, 0.159] <0.001 31% 0%
22 ForecastBench LLM Crowd (gpt-4o, claude-3.5-sonnet, gemini-1.5-pro) with news 0.241 0.089 0.039 0.067 0.165 0.154 [0.15, 0.158] <0.001 39% 76%
23 OpenAI GPT-4 (zero shot with freeze values) 0.222 0.116 0.047 0.086 0.169 0.154 [0.148, 0.161] <0.001 39% 0%
24 Anthropic Claude-3-5-Sonnet-20240620 (scratchpad with news) 0.189 0.155 0.081 0.123 0.172 0.156 [0.15, 0.161] <0.001 30% 0%
25 Qwen Qwen1.5-110B-Chat (scratchpad) 0.179 0.166 0.091 0.133 0.172 0.156 [0.151, 0.161] <0.001 29% 0%
26 ForecastBench LLM Crowd (gpt-4o, claude-3.5-sonnet, gemini-1.5-pro) with news 0.242 0.093 0.040 0.070 0.168 0.156 [0.152, 0.16] <0.001 39% 76%
27 ForecastBench LLM Crowd (gpt-4o, claude-3.5-sonnet, gemini-1.5-pro) with news 0.243 0.093 0.040 0.070 0.168 0.156 [0.152, 0.16] <0.001 38% 76%
28 Google Gemini-1.5-Pro (zero shot) 0.219 0.107 0.080 0.095 0.163 0.157 [0.151, 0.163] <0.001 38% 19%
29 Meta Llama-3-70b-Chat-Hf (zero shot with freeze values) 0.215 0.134 0.056 0.100 0.174 0.157 [0.151, 0.164] <0.001 33% 0%
30 Mistral AI Mistral-Large-Latest (scratchpad with freeze values) 0.201 0.143 0.075 0.113 0.172 0.157 [0.152, 0.163] <0.001 26% 0%
31 OpenAI GPT-4 (scratchpad) 0.197 0.157 0.072 0.120 0.177 0.158 [0.153, 0.163] <0.001 29% 0%
32 Anthropic Claude-3-Opus-20240229 (scratchpad with freeze values) 0.207 0.142 0.069 0.110 0.174 0.158 [0.152, 0.164] <0.001 27% 0%
33 Google Gemini-1.5-Pro (superforecaster with news 3) 0.187 0.164 0.087 0.130 0.175 0.159 [0.153, 0.165] <0.001 32% 0%
34 OpenAI GPT-4-Turbo-2024-04-09 (zero shot) 0.203 0.141 0.082 0.115 0.172 0.159 [0.153, 0.166] <0.001 32% 0%
35 Google Gemini-1.5-Flash (scratchpad with freeze values) 0.194 0.161 0.080 0.126 0.178 0.160 [0.153, 0.167] <0.001 32% 0%
36 Anthropic Claude-3-5-Sonnet-20240620 (superforecaster with news 3) 0.193 0.163 0.082 0.128 0.178 0.160 [0.154, 0.166] <0.001 29% 2%
37 OpenAI GPT-4-Turbo-2024-04-09 (superforecaster with news 3) 0.207 0.143 0.078 0.114 0.175 0.161 [0.155, 0.166] <0.001 28% 8%
38 OpenAI GPT-4-Turbo-2024-04-09 (scratchpad with news) 0.211 0.132 0.087 0.112 0.172 0.162 [0.156, 0.167] <0.001 28% 0%
39 Anthropic Claude-3-Opus-20240229 (zero shot) 0.189 0.170 0.091 0.135 0.179 0.162 [0.155, 0.169] <0.001 34% 0%
40 Qwen Qwen1.5-110B-Chat (scratchpad with news with freeze values) 0.207 0.157 0.074 0.121 0.182 0.164 [0.158, 0.17] <0.001 27% 0%
41 Meta Llama-3-70b-Chat-Hf (scratchpad with freeze values) 0.224 0.140 0.064 0.107 0.182 0.165 [0.16, 0.171] <0.001 26% 0%
42 Mistral AI Mixtral-8x22B-Instruct-V0.1 (scratchpad with freeze values) 0.213 0.147 0.084 0.119 0.180 0.166 [0.16, 0.173] <0.001 31% 0%
43 Anthropic Claude-3-5-Sonnet-20240620 (zero shot) 0.200 0.165 0.093 0.133 0.182 0.167 [0.159, 0.174] <0.001 34% 1%
44 OpenAI GPT-4o (superforecaster with news 3) 0.210 0.156 0.080 0.123 0.183 0.167 [0.161, 0.172] <0.001 28% 6%
45 Google Gemini-1.5-Flash (zero shot with freeze values) 0.237 0.127 0.060 0.097 0.182 0.167 [0.16, 0.175] <0.001 40% 0%
46 Anthropic Claude-2.1 (scratchpad) 0.236 0.124 0.067 0.099 0.180 0.168 [0.162, 0.173] <0.001 40% 24%
47 Mistral AI Mistral-Large-Latest (zero shot with freeze values) 0.215 0.160 0.071 0.121 0.187 0.168 [0.16, 0.176] <0.001 31% 0%
48 Google Gemini-1.5-Flash (scratchpad) 0.194 0.178 0.102 0.145 0.186 0.170 [0.163, 0.176] <0.001 28% 0%
49 OpenAI GPT-4o (superforecaster with news 1) 0.215 0.149 0.097 0.126 0.182 0.170 [0.163, 0.178] <0.001 32% 0%
50 OpenAI GPT-4-Turbo-2024-04-09 (superforecaster with news 1) 0.216 0.160 0.083 0.127 0.188 0.171 [0.165, 0.178] <0.001 30% 0%
51 Meta Llama-3-70b-Chat-Hf (zero shot) 0.215 0.156 0.092 0.128 0.185 0.171 [0.165, 0.177] <0.001 28% 0%
52 Anthropic Claude-3-5-Sonnet-20240620 (superforecaster with news 1) 0.212 0.166 0.090 0.133 0.189 0.172 [0.166, 0.179] <0.001 31% 16%
53 Qwen Qwen1.5-110B-Chat (zero shot with freeze values) 0.228 0.155 0.069 0.117 0.191 0.172 [0.165, 0.18] <0.001 32% 1%
54 Anthropic Claude-2.1 (scratchpad with freeze values) 0.236 0.148 0.058 0.109 0.192 0.173 [0.166, 0.179] <0.001 38% 19%
55 OpenAI GPT-4o (zero shot with freeze values) 0.229 0.143 0.080 0.116 0.186 0.173 [0.165, 0.18] <0.001 37% 1%
56 Anthropic Claude-3-Opus-20240229 (scratchpad) 0.207 0.179 0.091 0.140 0.193 0.173 [0.168, 0.179] <0.001 26% 0%
57 OpenAI GPT-4o (scratchpad with SECOND news) 0.236 0.142 0.076 0.113 0.189 0.174 [0.169, 0.18] <0.001 29% 8%
58 Qwen Qwen1.5-110B-Chat (scratchpad with news) 0.207 0.182 0.092 0.143 0.194 0.175 [0.169, 0.18] <0.001 26% 0%
59 Meta Llama-3-8b-Chat-Hf (zero shot with freeze values) 0.222 0.142 0.108 0.127 0.182 0.175 [0.167, 0.182] <0.001 43% 0%
60 Mistral AI Mistral-Large-Latest (scratchpad) 0.201 0.187 0.099 0.148 0.194 0.175 [0.169, 0.18] <0.001 25% 0%
61 OpenAI GPT-4o (zero shot) 0.229 0.161 0.069 0.121 0.195 0.175 [0.168, 0.182] <0.001 30% 3%
62 Google Gemini-1.5-Flash (scratchpad with news with freeze values) 0.221 0.164 0.089 0.131 0.192 0.176 [0.169, 0.183] <0.001 29% 0%
63 Anthropic Claude-3-5-Sonnet-20240620 (scratchpad with SECOND news) 0.224 0.167 0.084 0.131 0.196 0.177 [0.172, 0.183] <0.001 29% 1%
64 Mistral AI Mixtral-8x22B-Instruct-V0.1 (scratchpad) 0.213 0.184 0.095 0.145 0.198 0.179 [0.173, 0.185] <0.001 27% 0%
65 Anthropic Claude-2.1 (scratchpad with news) 0.241 0.158 0.071 0.120 0.200 0.180 [0.174, 0.187] <0.001 35% 18%
66 Qwen Qwen1.5-110B-Chat (superforecaster with news 1) 0.211 0.191 0.098 0.150 0.201 0.181 [0.174, 0.187] <0.001 31% 15%
67 Qwen Qwen1.5-110B-Chat (superforecaster with news 3) 0.216 0.190 0.090 0.146 0.203 0.181 [0.176, 0.187] <0.001 25% 3%
68 Anthropic Claude-2.1 (scratchpad with news with freeze values) 0.241 0.164 0.072 0.124 0.203 0.182 [0.176, 0.189] <0.001 34% 16%
69 Anthropic Claude-3-Opus-20240229 (superforecaster with news 1) 0.219 0.186 0.096 0.147 0.202 0.183 [0.176, 0.189] <0.001 28% 8%
70 Anthropic Claude-3-Opus-20240229 (scratchpad with news with freeze values) 0.224 0.180 0.091 0.141 0.202 0.183 [0.176, 0.189] <0.001 24% 0%
71 Google Gemini-1.5-Pro (superforecaster with news 1) 0.221 0.182 0.098 0.146 0.201 0.183 [0.176, 0.19] <0.001 30% 4%
72 OpenAI GPT-4 (zero shot) 0.222 0.187 0.091 0.145 0.204 0.184 [0.178, 0.189] <0.001 26% 0%
73 Mistral AI Mixtral-8x22B-Instruct-V0.1 (zero shot with freeze values) 0.219 0.189 0.095 0.148 0.204 0.184 [0.175, 0.192] <0.001 29% 0%
74 Mistral AI Mixtral-8x22B-Instruct-V0.1 (zero shot) 0.219 0.179 0.112 0.150 0.199 0.185 [0.178, 0.192] <0.001 27% 0%
75 Meta Llama-3-70b-Chat-Hf (scratchpad) 0.224 0.188 0.093 0.147 0.206 0.185 [0.18, 0.19] <0.001 23% 0%
76 Google Gemini-1.5-Flash (scratchpad with news) 0.221 0.185 0.108 0.151 0.203 0.186 [0.18, 0.192] <0.001 26% 0%
77 Google Gemini-1.5-Flash (superforecaster with news 2) 0.244 0.146 0.108 0.129 0.195 0.187 [0.18, 0.193] <0.001 29% 13%
78 Qwen Qwen1.5-110B-Chat (zero shot) 0.228 0.182 0.100 0.146 0.205 0.187 [0.181, 0.193] <0.001 27% 1%
79 Mistral AI Mixtral-8x22B-Instruct-V0.1 (scratchpad with news with freeze values) 0.233 0.187 0.088 0.144 0.210 0.189 [0.183, 0.195] <0.001 26% 0%
80 Anthropic Claude-3-Opus-20240229 (superforecaster with news 3) 0.219 0.200 0.106 0.159 0.209 0.189 [0.183, 0.195] <0.001 26% 5%
81 Meta Llama-3-8b-Chat-Hf (zero shot) 0.222 0.181 0.125 0.156 0.201 0.189 [0.181, 0.198] <0.001 41% 0%
82 Mistral AI Mixtral-8x22B-Instruct-V0.1 (superforecaster with news 1) 0.234 0.174 0.107 0.145 0.204 0.189 [0.182, 0.196] <0.001 30% 19%
83 Mistral AI Mistral-Large-Latest (zero shot) 0.215 0.195 0.125 0.164 0.205 0.190 [0.183, 0.197] <0.001 27% 0%
84 Anthropic Claude-3-Opus-20240229 (scratchpad with news) 0.224 0.197 0.102 0.156 0.211 0.190 [0.184, 0.196] <0.001 24% 0%
85 Anthropic Claude-3-Opus-20240229 (superforecaster with news 2) 0.228 0.193 0.108 0.156 0.210 0.192 [0.185, 0.199] <0.001 26% 1%
86 Mistral AI Mixtral-8x22B-Instruct-V0.1 (scratchpad with news) 0.233 0.195 0.093 0.150 0.214 0.192 [0.186, 0.197] <0.001 25% 0%
87 Mistral AI Mixtral-8x22B-Instruct-V0.1 (superforecaster with news 3) 0.232 0.196 0.096 0.152 0.214 0.192 [0.187, 0.197] <0.001 26% 11%
88 Anthropic Claude-3-5-Sonnet-20240620 (superforecaster with news 2) 0.226 0.201 0.103 0.158 0.214 0.192 [0.185, 0.199] <0.001 26% 1%
89 Mistral AI Mixtral-8x7B-Instruct-V0.1 (zero shot with freeze values) 0.264 0.153 0.092 0.126 0.209 0.195 [0.187, 0.204] <0.001 38% 0%
90 Google Gemini-1.5-Flash (superforecaster with news 3) 0.230 0.206 0.104 0.161 0.218 0.195 [0.189, 0.202] <0.001 25% 9%
91 Mistral AI Mixtral-8x7B-Instruct-V0.1 (scratchpad) 0.250 0.174 0.099 0.141 0.212 0.196 [0.189, 0.202] <0.001 29% 12%
92 Mistral AI Mixtral-8x7B-Instruct-V0.1 (scratchpad with freeze values) 0.250 0.176 0.097 0.141 0.213 0.196 [0.189, 0.203] <0.001 29% 11%
93 Anthropic Claude-2.1 (superforecaster with news 2) 0.268 0.164 0.071 0.124 0.216 0.196 [0.188, 0.203] <0.001 38% 25%
94 Google Gemini-1.5-Pro (superforecaster with news 2) 0.245 0.189 0.104 0.152 0.217 0.198 [0.19, 0.206] <0.001 30% 0%
95 Anthropic Claude-2.1 (zero shot with freeze values) 0.246 0.193 0.106 0.155 0.219 0.200 [0.192, 0.209] <0.001 33% 0%
96 OpenAI GPT-4o (superforecaster with news 2) 0.258 0.184 0.094 0.144 0.221 0.201 [0.194, 0.209] <0.001 28% 4%
97 OpenAI GPT-4-Turbo-2024-04-09 (superforecaster with news 2) 0.247 0.193 0.109 0.156 0.220 0.202 [0.194, 0.209] <0.001 28% 6%
98 Qwen Qwen1.5-110B-Chat (superforecaster with news 2) 0.240 0.212 0.102 0.164 0.226 0.202 [0.196, 0.208] <0.001 25% 4%
99 Mistral AI Mixtral-8x7B-Instruct-V0.1 (superforecaster with news 3) 0.274 0.153 0.101 0.130 0.213 0.202 [0.196, 0.208] <0.001 32% 20%
100 Anthropic Claude-2.1 (superforecaster with news 1) 0.266 0.180 0.087 0.139 0.223 0.203 [0.196, 0.21] <0.001 33% 25%
101 Google Gemini-1.5-Flash (zero shot) 0.237 0.196 0.141 0.172 0.217 0.205 [0.197, 0.213] <0.001 30% 1%
102 Mistral AI Mixtral-8x22B-Instruct-V0.1 (superforecaster with news 2) 0.248 0.208 0.103 0.162 0.228 0.205 [0.199, 0.211] <0.001 24% 2%
103 Mistral AI Mistral-Large-Latest (scratchpad with news with freeze values) 0.254 0.207 0.104 0.162 0.230 0.208 [0.201, 0.214] <0.001 22% 0%
104 Mistral AI Mistral-Large-Latest (superforecaster with news 1) 0.249 0.207 0.121 0.169 0.228 0.209 [0.202, 0.217] <0.001 26% 24%
105 Mistral AI Mistral-Large-Latest (scratchpad with news) 0.254 0.212 0.115 0.170 0.233 0.212 [0.206, 0.218] <0.001 22% 0%
106 Mistral AI Mistral-Large-Latest (superforecaster with news 2) 0.239 0.234 0.125 0.186 0.237 0.213 [0.205, 0.22] <0.001 26% 9%
107 Mistral AI Mixtral-8x7B-Instruct-V0.1 (scratchpad with news) 0.286 0.177 0.104 0.145 0.231 0.215 [0.209, 0.222] <0.001 29% 17%
108 Anthropic Claude-2.1 (superforecaster with news 3) 0.264 0.215 0.107 0.167 0.239 0.216 [0.209, 0.222] <0.001 28% 11%
109 Mistral AI Mixtral-8x7B-Instruct-V0.1 (scratchpad with news with freeze values) 0.286 0.174 0.111 0.147 0.230 0.216 [0.21, 0.223] <0.001 29% 16%
110 Anthropic Claude-2.1 (zero shot) 0.246 0.234 0.129 0.188 0.240 0.217 [0.21, 0.223] <0.001 26% 1%
111 Mistral AI Mixtral-8x7B-Instruct-V0.1 (zero shot) 0.264 0.206 0.125 0.170 0.235 0.217 [0.208, 0.227] <0.001 33% 1%
112 Mistral AI Mixtral-8x7B-Instruct-V0.1 (superforecaster with news 1) 0.301 0.164 0.096 0.135 0.233 0.218 [0.211, 0.225] <0.001 33% 20%
113 Google Gemini-1.5-Flash (superforecaster with news 1) 0.245 0.231 0.145 0.193 0.238 0.219 [0.211, 0.227] <0.001 25% 17%
114 ForecastBench Always 0.5 0.250 0.250 0.115 0.191 0.250 0.220 [0.218, 0.223] <0.001 25% 0%
115 Meta Llama-3-8b-Chat-Hf (scratchpad with freeze values) 0.274 0.211 0.115 0.169 0.242 0.221 [0.215, 0.228] <0.001 23% 0%
116 Meta Llama-2-70b-Chat-Hf (scratchpad with freeze values) 0.266 0.227 0.119 0.180 0.246 0.223 [0.217, 0.228] <0.001 23% 1%
117 Mistral AI Mistral-Large-Latest (superforecaster with news 3) 0.271 0.225 0.114 0.176 0.248 0.224 [0.217, 0.23] <0.001 22% 4%
118 Anthropic Claude-3-Haiku-20240307 (zero shot with freeze values) 0.284 0.209 0.105 0.163 0.246 0.224 [0.217, 0.23] <0.001 24% 1%
119 Meta Llama-2-70b-Chat-Hf (scratchpad) 0.266 0.236 0.114 0.183 0.251 0.224 [0.219, 0.23] <0.001 22% 1%
120 Mistral AI Mixtral-8x7B-Instruct-V0.1 (superforecaster with news 2) 0.295 0.186 0.118 0.156 0.240 0.225 [0.218, 0.232] <0.001 32% 20%
121 Anthropic Claude-3-Haiku-20240307 (scratchpad with freeze values) 0.280 0.228 0.121 0.181 0.254 0.231 [0.225, 0.236] <0.001 20% 1%
122 Anthropic Claude-3-Haiku-20240307 (superforecaster with news 2) 0.269 0.262 0.119 0.200 0.265 0.234 [0.228, 0.24] <0.001 22% 6%
123 Meta Llama-3-8b-Chat-Hf (scratchpad) 0.274 0.248 0.128 0.195 0.261 0.235 [0.228, 0.241] <0.001 22% 0%
124 ForecastBench Always 0 0.269 0.203 0.197 0.200 0.236 0.235 [0.222, 0.248] <0.001 65% 0%
125 Anthropic Claude-3-Haiku-20240307 (zero shot) 0.284 0.238 0.123 0.188 0.261 0.236 [0.23, 0.241] <0.001 24% 1%
126 Anthropic Claude-3-Haiku-20240307 (scratchpad) 0.280 0.261 0.129 0.203 0.271 0.242 [0.236, 0.247] <0.001 20% 1%
127 Anthropic Claude-3-Haiku-20240307 (scratchpad with news with freeze values) 0.305 0.243 0.120 0.189 0.274 0.247 [0.241, 0.253] <0.001 20% 1%
128 Anthropic Claude-3-Haiku-20240307 (superforecaster with news 3) 0.287 0.278 0.136 0.216 0.282 0.251 [0.246, 0.257] <0.001 22% 16%
129 Anthropic Claude-3-Haiku-20240307 (scratchpad with news) 0.305 0.263 0.128 0.204 0.284 0.254 [0.249, 0.26] <0.001 20% 1%
130 OpenAI GPT-3.5-Turbo-0125 (scratchpad with freeze values) 0.296 0.271 0.145 0.216 0.284 0.256 [0.25, 0.262] <0.001 20% 0%
131 OpenAI GPT-3.5-Turbo-0125 (scratchpad) 0.296 0.280 0.140 0.219 0.288 0.257 [0.251, 0.263] <0.001 20% 0%
132 Meta Llama-2-70b-Chat-Hf (zero shot with freeze values) 0.310 0.282 0.156 0.227 0.296 0.269 [0.259, 0.278] <0.001 27% 1%
133 Anthropic Claude-3-Haiku-20240307 (superforecaster with news 1) 0.300 0.323 0.152 0.248 0.312 0.274 [0.267, 0.282] <0.001 22% 6%
134 Meta Llama-2-70b-Chat-Hf (zero shot) 0.310 0.321 0.165 0.253 0.316 0.282 [0.274, 0.289] <0.001 25% 1%
135 ForecastBench Random Uniform 0.342 0.321 0.195 0.266 0.332 0.304 [0.294, 0.314] <0.001 28% 0%
136 OpenAI GPT-3.5-Turbo-0125 (zero shot with freeze values) 0.446 0.243 0.151 0.203 0.344 0.324 [0.314, 0.335] <0.001 28% 0%
137 OpenAI GPT-3.5-Turbo-0125 (zero shot) 0.446 0.338 0.218 0.285 0.392 0.366 [0.357, 0.375] <0.001 21% 0%
138 ForecastBench Always 1 0.731 0.797 0.532 0.681 0.764 0.706 [0.691, 0.72] <0.001 22% 0%