Atlas

Models

← All models

GPT-4 (0314)

OpenAI · 2023-03-14

The original March 14, 2023 GPT-4 API snapshot. OpenAI lists gpt-4-0314 separately from later GPT-4 snapshots and the mutable gpt-4 alias. METR's Time Horizon 1.1 raw runs identify this checkpoint as GPT-4 0314.

Benchmark scores

BenchmarkScore
AidanBench909
Arena Text — Business, Management & Financial Ops (No Style Control)1145
Arena Text — Chinese (No Style Control)1184
Arena Text — Creative Writing (No Style Control)1190
Arena Text — English (No Style Control)1220
Arena Text — Entertainment, Sports & Media (No Style Control)1184
Arena Text — Exclude Ties (No Style Control)1086
Arena Text — Exclude Ties (Style Controlled)1202
Arena Text — Expert (No Style Control)1149
Arena Text — French (No Style Control)1220
Arena Text — French (Style Controlled)1308
Arena Text — German (No Style Control)1198
Arena Text — Hard Prompts (No Style Control)1200
Arena Text — Hard Prompts English (No Style Control)1209
Arena Text — Instruction Following (No Style Control)1201
Arena Text — Japanese (No Style Control)1137
Arena Text — Korean (No Style Control)1087
Arena Text — Legal & Government (No Style Control)1192
Arena Text — Life, Physical & Social Science (No Style Control)1177
Arena Text — Longer Query (No Style Control)1190
Arena Text — Medicine & Healthcare (No Style Control)1145
Arena Text — Multi-turn (No Style Control)1206
Arena Text — Non-English (No Style Control)1194
Arena Text — Non-English (Style Controlled)1272
Arena Text — Occupational Mathematical (No Style Control)1228
Arena Text — Polish (No Style Control)1265
Arena Text — Polish (Style Controlled)1337
Arena Text — Russian (No Style Control)1195
Arena Text — Software & IT Services (No Style Control)1195
Arena Text — Spanish (No Style Control)1197
Loading Atlas data…