GPT-4.5
GPT-4.5 is a closed AI model from OpenAI with an index score of 125 (GPT-4 = 100), rank 145 of 217.
125109–141Index score#145 of 217 models
–Price per million tokensno public API price found
–Context windowtokens it reads at once
27 Feb 2025ReleasedReplaced by GPT-5
Where it sitson the frontier
o1 reached this score 3 months earlier.
Test resultsby area · tick = best by any model
Reasoning & knowledgeevidence
Mathsevidence
Codingevidence
Agentsno evidence yet
Professional workno evidence yet
Novel problemsevidence
Long tasksno evidence yet
Third-party results only, at the model’s best standard setting. Hover a bar for who ran it.
What others sayrank on 9 leaderboards
The Model Indexthird-party tests, our method#145of 217 · 125
LMArena · Creative writingcreative writing#53of 360 · 1436LMArena · Instruction followingfollowing detailed instructions#72of 362 · 1437LMArena · Textoverall chat quality, judged by people#76of 362 · 1445LMArena · Long promptslong, complex requests#87of 340 · 1443LMArena · Codingcoding questions in chat#123of 357 · 1458LMArena · Mathsmaths questions in chat#125of 346 · 1407LMArena · Visionunderstanding images#66of 145 · 1225Epoch Capabilities Indexoverall capability from many benchmarks#125of 231 · 136.5Best attop three among current models
Replaced by GPT-5; it no longer leads any category among current models.
Where to use itprices checked 18 Nov 2026
Host prices from OpenRouter; going direct can differ.
Facts and sources
Released27 Feb 2025
Announcementen.wikipedia.org ↗
Release entryGPT-4.5
Scored asGPT-4.5
Output speed–
Hugging Faceclosed weights