Human Leaderboard: overall

Updated Nov. 21, 2024

  • 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=316) Market Score (resolved) (N=20) Market Score (unresolved) (N=57) Market Score (overall) (N=77) Overall Resolved Score (N=336) Overall Score (N=393) Overall Score 95% CI Pairwise p-value comparing to No. 1 (bootstrapped) Pct. more accurate than No. 1 Pct. Imputed
1 ForecastBench Superforecaster median forecast 0.126 0.084 0.051 0.060 0.105 0.093 [0.073, 0.112] 0% 0%
2 Anthropic Claude-3-5-Sonnet-20240620 (scratchpad with freeze values) 0.139 0.177 0.056 0.087 0.158 0.113 [0.091, 0.135] 0.001 34% 0%
3 ForecastBench Public median forecast 0.156 0.121 0.053 0.071 0.139 0.114 [0.094, 0.133] <0.001 25% 0%
4 Anthropic Claude-3-5-Sonnet-20240620 (scratchpad with news with freeze values) 0.144 0.160 0.069 0.093 0.152 0.118 [0.097, 0.14] <0.001 31% 0%
5 OpenAI GPT-4 (zero shot with freeze values) 0.164 0.129 0.059 0.077 0.147 0.121 [0.099, 0.142] <0.001 33% 0%
6 Anthropic Claude-3-5-Sonnet-20240620 (zero shot with freeze values) 0.145 0.202 0.061 0.098 0.173 0.121 [0.094, 0.148] <0.001 33% 0%
7 OpenAI GPT-4-Turbo-2024-04-09 (zero shot with freeze values) 0.170 0.129 0.057 0.076 0.150 0.123 [0.102, 0.144] <0.001 33% 0%
8 OpenAI GPT-4o (scratchpad with news with freeze values) 0.168 0.100 0.074 0.081 0.134 0.125 [0.106, 0.143] <0.001 27% 0%
9 Anthropic Claude-3-5-Sonnet-20240620 (scratchpad) 0.139 0.244 0.068 0.114 0.192 0.126 [0.105, 0.148] <0.001 29% 0%
10 OpenAI GPT-4o (scratchpad with freeze values) 0.163 0.172 0.063 0.092 0.168 0.127 [0.107, 0.147] <0.001 28% 0%
11 Anthropic Claude-3-5-Sonnet-20240620 (scratchpad with news) 0.144 0.184 0.091 0.115 0.164 0.130 [0.109, 0.15] <0.001 26% 0%
12 Anthropic Claude-3-5-Sonnet-20240620 (superforecaster with news 3) 0.159 0.154 0.083 0.102 0.157 0.131 [0.112, 0.149] <0.001 26% 3%
13 Mistral AI Mistral-Large-Latest (zero shot with freeze values) 0.170 0.195 0.061 0.096 0.182 0.133 [0.11, 0.155] <0.001 25% 0%
14 Anthropic Claude-3-5-Sonnet-20240620 (superforecaster with news 1) 0.153 0.178 0.089 0.112 0.166 0.133 [0.112, 0.153] <0.001 26% 0%
15 OpenAI GPT-4o (scratchpad) 0.163 0.175 0.078 0.103 0.169 0.133 [0.116, 0.15] <0.001 25% 0%
16 Anthropic Claude-3-Opus-20240229 (zero shot with freeze values) 0.173 0.183 0.062 0.094 0.178 0.133 [0.109, 0.157] <0.001 27% 0%
17 Anthropic Claude-3-Opus-20240229 (scratchpad with freeze values) 0.163 0.163 0.086 0.106 0.163 0.135 [0.117, 0.152] <0.001 24% 0%
18 Mistral AI Mistral-Large-Latest (scratchpad with freeze values) 0.161 0.153 0.095 0.110 0.157 0.136 [0.119, 0.152] <0.001 23% 0%
19 OpenAI GPT-4o (scratchpad with news) 0.168 0.171 0.082 0.105 0.169 0.137 [0.116, 0.157] <0.001 23% 0%
20 OpenAI GPT-4-Turbo-2024-04-09 (scratchpad with freeze values) 0.179 0.149 0.086 0.102 0.164 0.140 [0.118, 0.163] <0.001 28% 0%
21 Anthropic Claude-3-5-Sonnet-20240620 (zero shot) 0.145 0.288 0.084 0.137 0.216 0.141 [0.114, 0.167] <0.001 25% 0%
22 Google Gemini-1.5-Pro (scratchpad with news with freeze values) 0.172 0.153 0.096 0.111 0.163 0.142 [0.123, 0.16] <0.001 23% 0%
23 Google Gemini-1.5-Pro (scratchpad with freeze values) 0.168 0.210 0.082 0.115 0.189 0.142 [0.124, 0.159] <0.001 24% 0%
24 Google Gemini-1.5-Pro (scratchpad) 0.168 0.206 0.084 0.115 0.187 0.142 [0.124, 0.159] <0.001 23% 0%
25 OpenAI GPT-4 (scratchpad with freeze values) 0.178 0.165 0.084 0.105 0.172 0.142 [0.123, 0.161] <0.001 22% 0%
26 Meta Llama-3-70b-Chat-Hf (scratchpad with freeze values) 0.189 0.133 0.081 0.095 0.161 0.142 [0.126, 0.158] <0.001 25% 0%
27 Meta Llama-3-70b-Chat-Hf (zero shot with freeze values) 0.181 0.203 0.068 0.103 0.192 0.142 [0.118, 0.167] <0.001 27% 0%
28 OpenAI GPT-4-Turbo-2024-04-09 (zero shot) 0.170 0.221 0.079 0.116 0.196 0.143 [0.124, 0.163] <0.001 22% 0%
29 Anthropic Claude-2.1 (scratchpad with freeze values) 0.213 0.043 0.086 0.075 0.128 0.144 [0.127, 0.161] <0.001 27% 23%
30 Anthropic Claude-3-Opus-20240229 (scratchpad) 0.163 0.194 0.101 0.125 0.179 0.144 [0.127, 0.161] <0.001 22% 0%
31 Anthropic Claude-3-Opus-20240229 (superforecaster with news 1) 0.161 0.210 0.098 0.127 0.185 0.144 [0.124, 0.165] <0.001 23% 0%
32 Google Gemini-1.5-Flash (zero shot with freeze values) 0.183 0.234 0.060 0.105 0.208 0.144 [0.117, 0.171] <0.001 25% 0%
33 OpenAI GPT-4o (zero shot with freeze values) 0.201 0.183 0.056 0.089 0.192 0.145 [0.121, 0.169] <0.001 30% 0%
34 Google Gemini-1.5-Pro (zero shot with freeze values) 0.186 0.227 0.061 0.104 0.206 0.145 [0.119, 0.171] <0.001 26% 0%
35 Qwen Qwen1.5-110B-Chat (zero shot with freeze values) 0.195 0.173 0.068 0.095 0.184 0.145 [0.124, 0.166] <0.001 23% 0%
36 Google Gemini-1.5-Pro (scratchpad with news) 0.172 0.150 0.107 0.118 0.161 0.145 [0.127, 0.164] <0.001 23% 0%
37 OpenAI GPT-4-Turbo-2024-04-09 (scratchpad) 0.179 0.197 0.082 0.112 0.188 0.145 [0.129, 0.161] <0.001 23% 0%
38 Mistral AI Mixtral-8x22B-Instruct-V0.1 (scratchpad with freeze values) 0.183 0.192 0.078 0.108 0.188 0.145 [0.129, 0.162] <0.001 21% 0%
39 OpenAI GPT-4-Turbo-2024-04-09 (scratchpad with news with freeze values) 0.188 0.125 0.101 0.107 0.156 0.148 [0.126, 0.17] <0.001 27% 0%
40 OpenAI GPT-4 (scratchpad) 0.178 0.173 0.097 0.117 0.176 0.148 [0.133, 0.162] <0.001 18% 0%
41 Mistral AI Mixtral-8x22B-Instruct-V0.1 (zero shot with freeze values) 0.189 0.179 0.086 0.110 0.184 0.150 [0.124, 0.175] <0.001 27% 0%
42 Anthropic Claude-2.1 (scratchpad) 0.213 0.108 0.084 0.090 0.161 0.152 [0.133, 0.17] <0.001 24% 23%
43 ForecastBench Imputed Forecaster 0.250 0.076 0.047 0.055 0.163 0.152 [0.135, 0.169] <0.001 29% 100%
44 Mistral AI Mistral-Large-Latest (scratchpad) 0.161 0.223 0.116 0.144 0.192 0.152 [0.135, 0.17] <0.001 22% 0%
45 OpenAI GPT-4 (zero shot) 0.164 0.177 0.128 0.141 0.171 0.152 [0.135, 0.17] <0.001 24% 0%
46 Anthropic Claude-2.1 (zero shot with freeze values) 0.220 0.169 0.060 0.088 0.195 0.154 [0.132, 0.176] <0.001 28% 0%
47 Qwen Qwen1.5-110B-Chat (scratchpad with freeze values) 0.193 0.193 0.089 0.116 0.193 0.154 [0.137, 0.172] <0.001 19% 0%
48 Qwen Qwen1.5-110B-Chat (scratchpad with news with freeze values) 0.185 0.189 0.101 0.124 0.187 0.155 [0.136, 0.173] <0.001 23% 0%
49 OpenAI GPT-4-Turbo-2024-04-09 (scratchpad with news) 0.188 0.172 0.106 0.123 0.180 0.155 [0.135, 0.176] <0.001 25% 0%
50 Mistral AI Mixtral-8x22B-Instruct-V0.1 (scratchpad with news with freeze values) 0.197 0.157 0.099 0.114 0.177 0.156 [0.137, 0.174] <0.001 23% 0%
51 Google Gemini-1.5-Pro (superforecaster with news 3) 0.189 0.146 0.116 0.124 0.167 0.156 [0.138, 0.175] <0.001 22% 0%
52 Mistral AI Mixtral-8x22B-Instruct-V0.1 (scratchpad with news) 0.197 0.174 0.097 0.117 0.186 0.157 [0.14, 0.174] <0.001 22% 0%
53 Anthropic Claude-3-5-Sonnet-20240620 (scratchpad with SECOND news) 0.204 0.159 0.094 0.111 0.181 0.157 [0.138, 0.176] <0.001 18% 0%
54 OpenAI GPT-4o (zero shot) 0.201 0.197 0.086 0.115 0.199 0.158 [0.138, 0.178] <0.001 22% 0%
55 OpenAI GPT-4o (superforecaster with news 3) 0.203 0.151 0.100 0.113 0.177 0.158 [0.139, 0.177] <0.001 21% 6%
56 Anthropic Claude-3-Opus-20240229 (zero shot) 0.173 0.258 0.105 0.145 0.215 0.159 [0.134, 0.183] <0.001 19% 0%
57 Google Gemini-1.5-Flash (scratchpad with freeze values) 0.206 0.174 0.091 0.112 0.190 0.159 [0.138, 0.18] <0.001 20% 0%
58 Mistral AI Mixtral-8x22B-Instruct-V0.1 (scratchpad) 0.183 0.206 0.112 0.137 0.195 0.160 [0.143, 0.177] <0.001 21% 0%
59 Anthropic Claude-3-Opus-20240229 (superforecaster with news 3) 0.178 0.141 0.144 0.143 0.159 0.161 [0.14, 0.181] <0.001 22% 6%
60 Qwen Qwen1.5-110B-Chat (scratchpad with news) 0.185 0.201 0.114 0.137 0.193 0.161 [0.142, 0.18] <0.001 22% 0%
61 Mistral AI Mistral-Large-Latest (zero shot) 0.170 0.213 0.134 0.154 0.191 0.162 [0.138, 0.186] <0.001 21% 0%
62 Anthropic Claude-3-5-Sonnet-20240620 (superforecaster with news 2) 0.193 0.165 0.119 0.131 0.179 0.162 [0.14, 0.184] <0.001 22% 0%
63 Meta Llama-3-8b-Chat-Hf (zero shot with freeze values) 0.205 0.202 0.090 0.119 0.204 0.162 [0.138, 0.187] <0.001 24% 0%
64 Mistral AI Mixtral-8x7B-Instruct-V0.1 (zero shot with freeze values) 0.214 0.230 0.071 0.113 0.222 0.163 [0.134, 0.193] <0.001 32% 0%
65 Anthropic Claude-3-Opus-20240229 (scratchpad with news with freeze values) 0.190 0.180 0.123 0.138 0.185 0.164 [0.143, 0.184] <0.001 24% 0%
66 Mistral AI Mixtral-8x7B-Instruct-V0.1 (scratchpad) 0.202 0.110 0.131 0.126 0.156 0.164 [0.144, 0.183] <0.001 27% 14%
67 Qwen Qwen1.5-110B-Chat (superforecaster with news 1) 0.201 0.232 0.091 0.128 0.216 0.164 [0.141, 0.187] <0.001 22% 0%
68 OpenAI GPT-4-Turbo-2024-04-09 (superforecaster with news 3) 0.205 0.151 0.115 0.124 0.178 0.164 [0.144, 0.185] <0.001 20% 11%
69 Qwen Qwen1.5-110B-Chat (scratchpad) 0.193 0.223 0.106 0.137 0.208 0.165 [0.147, 0.183] <0.001 20% 0%
70 Google Gemini-1.5-Flash (scratchpad with news with freeze values) 0.212 0.188 0.094 0.118 0.200 0.165 [0.143, 0.187] <0.001 22% 0%
71 Meta Llama-3-70b-Chat-Hf (zero shot) 0.181 0.233 0.119 0.149 0.207 0.165 [0.144, 0.186] <0.001 22% 0%
72 ForecastBench Always 0.5 0.250 0.175 0.047 0.081 0.213 0.165 [0.15, 0.18] <0.001 28% 53%
73 Google Gemini-1.5-Pro (zero shot) 0.186 0.275 0.101 0.146 0.230 0.166 [0.139, 0.193] <0.001 22% 0%
74 Meta Llama-3-8b-Chat-Hf (scratchpad with freeze values) 0.222 0.148 0.097 0.110 0.185 0.166 [0.152, 0.181] <0.001 23% 0%
75 Anthropic Claude-3-Opus-20240229 (scratchpad with news) 0.190 0.154 0.141 0.144 0.172 0.167 [0.146, 0.187] <0.001 23% 0%
76 Mistral AI Mistral-Large-Latest (scratchpad with news with freeze values) 0.202 0.171 0.118 0.132 0.187 0.167 [0.147, 0.187] <0.001 23% 0%
77 Qwen Qwen1.5-110B-Chat (zero shot) 0.195 0.217 0.113 0.140 0.206 0.167 [0.148, 0.186] <0.001 16% 1%
78 OpenAI GPT-4o (superforecaster with news 1) 0.199 0.215 0.108 0.136 0.207 0.168 [0.142, 0.193] <0.001 25% 0%
79 Anthropic Claude-2.1 (scratchpad with news with freeze values) 0.213 0.130 0.121 0.123 0.171 0.168 [0.147, 0.189] <0.001 22% 3%
80 Anthropic Claude-3-Opus-20240229 (superforecaster with news 2) 0.179 0.212 0.139 0.158 0.196 0.169 [0.145, 0.192] <0.001 22% 0%
81 Mistral AI Mixtral-8x22B-Instruct-V0.1 (superforecaster with news 1) 0.206 0.201 0.109 0.133 0.203 0.170 [0.146, 0.194] <0.001 21% 0%
82 ForecastBench LLM Crowd (gpt-4o, claude-3.5-sonnet, gemini-1.5-pro) with news 0.232 0.165 0.087 0.107 0.199 0.170 [0.153, 0.187] <0.001 16% 38%
83 ForecastBench LLM Crowd (gpt-4o, claude-3.5-sonnet, gemini-1.5-pro) with news 0.233 0.166 0.087 0.108 0.199 0.170 [0.153, 0.187] <0.001 17% 38%
84 ForecastBench LLM Crowd (gpt-4o, claude-3.5-sonnet, gemini-1.5-pro) with news 0.233 0.161 0.089 0.108 0.197 0.170 [0.153, 0.188] <0.001 18% 38%
85 Mistral AI Mixtral-8x22B-Instruct-V0.1 (superforecaster with news 3) 0.218 0.147 0.116 0.124 0.183 0.171 [0.154, 0.188] <0.001 16% 17%
86 Google Gemini-1.5-Flash (scratchpad) 0.206 0.195 0.116 0.137 0.200 0.171 [0.152, 0.191] <0.001 18% 0%
87 OpenAI GPT-4o (scratchpad with SECOND news) 0.237 0.142 0.093 0.106 0.189 0.172 [0.154, 0.189] <0.001 17% 2%
88 Google Gemini-1.5-Pro (superforecaster with news 1) 0.201 0.229 0.112 0.143 0.215 0.172 [0.149, 0.194] <0.001 23% 0%
89 Anthropic Claude-2.1 (scratchpad with news) 0.213 0.191 0.110 0.131 0.202 0.172 [0.15, 0.194] <0.001 22% 8%
90 Meta Llama-3-8b-Chat-Hf (zero shot) 0.205 0.259 0.097 0.139 0.232 0.172 [0.147, 0.198] <0.001 24% 0%
91 Google Gemini-1.5-Flash (scratchpad with news) 0.212 0.182 0.118 0.135 0.197 0.173 [0.152, 0.195] <0.001 21% 0%
92 Google Gemini-1.5-Flash (zero shot) 0.183 0.230 0.141 0.164 0.207 0.174 [0.148, 0.199] <0.001 20% 0%
93 OpenAI GPT-4-Turbo-2024-04-09 (superforecaster with news 1) 0.197 0.220 0.126 0.150 0.208 0.174 [0.15, 0.197] <0.001 21% 0%
94 Mistral AI Mistral-Large-Latest (scratchpad with news) 0.202 0.154 0.143 0.145 0.178 0.174 [0.154, 0.194] <0.001 21% 0%
95 Meta Llama-3-70b-Chat-Hf (scratchpad) 0.189 0.199 0.147 0.161 0.194 0.175 [0.159, 0.191] <0.001 24% 0%
96 Mistral AI Mistral-Large-Latest (superforecaster with news 1) 0.208 0.241 0.113 0.146 0.225 0.177 [0.152, 0.202] <0.001 23% 0%
97 OpenAI GPT-4o (superforecaster with news 2) 0.233 0.177 0.104 0.123 0.205 0.178 [0.154, 0.202] <0.001 23% 1%
98 Mistral AI Mixtral-8x22B-Instruct-V0.1 (zero shot) 0.189 0.282 0.128 0.168 0.235 0.178 [0.152, 0.205] <0.001 20% 0%
99 Mistral AI Mixtral-8x7B-Instruct-V0.1 (zero shot) 0.214 0.240 0.113 0.146 0.227 0.180 [0.152, 0.209] <0.001 22% 0%
100 Mistral AI Mistral-Large-Latest (superforecaster with news 2) 0.205 0.225 0.135 0.159 0.215 0.182 [0.158, 0.206] <0.001 22% 1%
101 Qwen Qwen1.5-110B-Chat (superforecaster with news 3) 0.219 0.201 0.127 0.146 0.210 0.182 [0.163, 0.201] <0.001 21% 6%
102 Anthropic Claude-2.1 (zero shot) 0.220 0.229 0.119 0.147 0.224 0.184 [0.164, 0.204] <0.001 19% 0%
103 Mistral AI Mixtral-8x22B-Instruct-V0.1 (superforecaster with news 2) 0.233 0.179 0.120 0.136 0.206 0.184 [0.166, 0.203] <0.001 23% 1%
104 Anthropic Claude-2.1 (superforecaster with news 3) 0.223 0.142 0.147 0.146 0.182 0.184 [0.164, 0.205] <0.001 23% 5%
105 Mistral AI Mixtral-8x7B-Instruct-V0.1 (superforecaster with news 2) 0.252 0.176 0.096 0.117 0.214 0.184 [0.16, 0.209] <0.001 30% 22%
106 Mistral AI Mistral-Large-Latest (superforecaster with news 3) 0.231 0.150 0.135 0.139 0.191 0.185 [0.166, 0.204] <0.001 21% 7%
107 Mistral AI Mixtral-8x7B-Instruct-V0.1 (superforecaster with news 1) 0.241 0.167 0.117 0.130 0.204 0.185 [0.161, 0.21] <0.001 27% 15%
108 Meta Llama-3-8b-Chat-Hf (scratchpad) 0.222 0.224 0.123 0.149 0.223 0.186 [0.169, 0.202] <0.001 22% 0%
109 OpenAI GPT-4-Turbo-2024-04-09 (superforecaster with news 2) 0.223 0.190 0.140 0.153 0.207 0.188 [0.164, 0.212] <0.001 25% 1%
110 Mistral AI Mixtral-8x7B-Instruct-V0.1 (scratchpad with freeze values) 0.202 0.202 0.166 0.175 0.202 0.188 [0.161, 0.216] <0.001 24% 11%
111 Meta Llama-2-70b-Chat-Hf (zero shot with freeze values) 0.228 0.174 0.144 0.151 0.201 0.189 [0.164, 0.215] <0.001 26% 0%
112 Google Gemini-1.5-Flash (superforecaster with news 2) 0.219 0.214 0.145 0.163 0.217 0.191 [0.166, 0.216] <0.001 20% 0%
113 Meta Llama-2-70b-Chat-Hf (scratchpad with freeze values) 0.221 0.252 0.131 0.162 0.236 0.191 [0.176, 0.207] <0.001 23% 0%
114 Google Gemini-1.5-Flash (superforecaster with news 3) 0.236 0.175 0.138 0.148 0.206 0.192 [0.17, 0.214] <0.001 19% 9%
115 Qwen Qwen1.5-110B-Chat (superforecaster with news 2) 0.224 0.212 0.145 0.162 0.218 0.193 [0.173, 0.213] <0.001 22% 3%
116 Google Gemini-1.5-Pro (superforecaster with news 2) 0.227 0.227 0.142 0.164 0.227 0.195 [0.167, 0.224] <0.001 21% 0%
117 Google Gemini-1.5-Flash (superforecaster with news 1) 0.225 0.273 0.130 0.167 0.249 0.196 [0.168, 0.223] <0.001 22% 0%
118 Anthropic Claude-3-Haiku-20240307 (scratchpad with freeze values) 0.237 0.183 0.148 0.157 0.210 0.197 [0.179, 0.215] <0.001 21% 0%
119 Mistral AI Mixtral-8x7B-Instruct-V0.1 (superforecaster with news 3) 0.246 0.136 0.156 0.150 0.191 0.198 [0.177, 0.219] <0.001 24% 13%
120 Anthropic Claude-3-Haiku-20240307 (superforecaster with news 2) 0.235 0.182 0.155 0.162 0.208 0.198 [0.181, 0.216] <0.001 20% 0%
121 Anthropic Claude-3-Haiku-20240307 (zero shot with freeze values) 0.292 0.145 0.095 0.108 0.219 0.200 [0.182, 0.217] <0.001 21% 0%
122 Mistral AI Mixtral-8x7B-Instruct-V0.1 (scratchpad with news with freeze values) 0.294 0.141 0.093 0.106 0.218 0.200 [0.18, 0.22] <0.001 24% 13%
123 Anthropic Claude-3-Haiku-20240307 (scratchpad) 0.237 0.206 0.152 0.166 0.222 0.202 [0.184, 0.219] <0.001 21% 0%
124 Meta Llama-2-70b-Chat-Hf (scratchpad) 0.221 0.289 0.155 0.190 0.255 0.205 [0.186, 0.224] <0.001 22% 0%
125 Anthropic Claude-2.1 (superforecaster with news 2) 0.239 0.186 0.167 0.172 0.213 0.205 [0.181, 0.23] <0.001 24% 11%
126 ForecastBench Random Uniform 0.301 0.307 0.047 0.115 0.304 0.208 [0.178, 0.238] <0.001 32% 53%
127 Anthropic Claude-3-Haiku-20240307 (scratchpad with news with freeze values) 0.271 0.171 0.142 0.150 0.221 0.210 [0.194, 0.227] <0.001 22% 0%
128 OpenAI GPT-3.5-Turbo-0125 (scratchpad with freeze values) 0.254 0.244 0.140 0.167 0.249 0.210 [0.192, 0.229] <0.001 22% 0%
129 Anthropic Claude-2.1 (superforecaster with news 1) 0.263 0.219 0.136 0.158 0.241 0.210 [0.187, 0.234] <0.001 23% 4%
130 Anthropic Claude-3-Haiku-20240307 (scratchpad with news) 0.271 0.187 0.143 0.155 0.229 0.213 [0.196, 0.23] <0.001 20% 0%
131 Mistral AI Mixtral-8x7B-Instruct-V0.1 (scratchpad with news) 0.294 0.160 0.129 0.137 0.227 0.216 [0.192, 0.239] <0.001 23% 13%
132 OpenAI GPT-3.5-Turbo-0125 (scratchpad) 0.254 0.268 0.150 0.181 0.261 0.217 [0.198, 0.236] <0.001 20% 0%
133 Anthropic Claude-3-Haiku-20240307 (zero shot) 0.292 0.183 0.132 0.145 0.237 0.219 [0.2, 0.237] <0.001 19% 0%
134 ForecastBench Always 0 0.321 0.350 0.047 0.126 0.336 0.224 [0.183, 0.264] <0.001 37% 53%
135 Anthropic Claude-3-Haiku-20240307 (superforecaster with news 3) 0.266 0.199 0.181 0.186 0.232 0.226 [0.208, 0.244] <0.001 19% 22%
136 Meta Llama-2-70b-Chat-Hf (zero shot) 0.228 0.275 0.206 0.224 0.251 0.226 [0.199, 0.252] <0.001 23% 1%
137 Anthropic Claude-3-Haiku-20240307 (superforecaster with news 1) 0.278 0.267 0.183 0.205 0.273 0.242 [0.216, 0.267] <0.001 22% 0%
138 OpenAI GPT-3.5-Turbo-0125 (zero shot with freeze values) 0.408 0.129 0.103 0.110 0.268 0.259 [0.23, 0.287] <0.001 31% 0%
139 OpenAI GPT-3.5-Turbo-0125 (zero shot) 0.408 0.259 0.182 0.202 0.333 0.305 [0.275, 0.335] <0.001 23% 0%
140 ForecastBench Always 1 0.511 0.350 0.047 0.126 0.431 0.319 [0.277, 0.36] <0.001 36% 53%