GPT-5.1

Nov 2025
Retires Apr 1, 2027Move toGPT-6.1 SolNewer version
GPT-5.1
Compare in
on
—per task?
$1.25 / $10per 1M in / outcache $0.13
128 tok/sfirst token in 38 s?
400Kcontext

Rankingat high effort?

Full ranking →
1Fable 5.11,520.82Opus 51,520.73Muse Spark 1.31,497.44Fable 51,4915Muse Spark 1.21,473
···
40Sonnet 4.51,401.841GPT-5.11,400.542MiMo v2 Pro1,399.2

Strong at

A
Science#55 of 312Long context#42 of 273Knowledge#49 of 318Instructions#34 of 245
B
Tool calling#72 of 242Reasoning#88 of 334Vision#43 of 124
C
Agents & work tasks#97 of 175

Alternatives?

ModelScoreCostPer taskWhat changes
CheaperMuse Spark 1.3maxwins 64%-42%$1.60Only auto tool_choice is supported+5BetterGPT-5.6 Lunaxhighwins 52%-87%$0.085OpenGLM 5.3maxwins 59%-37%$2.01No image input+4CheaperMuse Spark 1.2xhighwins 60%-42%$0.97Only auto tool_choice is supported+1CheaperGemini 3.8 Flashhighwins 59%-56%*$1.24Sampling parameters are ignored+1

History

17 versions
1,2001,3001,400Jul 24Jan 25Jul 25Jan 26Release dateArena Text · Korean4o4.1 · Apr5 · Aug5.1 · Nov5.2 · Dec5.4 · Mar5.5
1,2001,3001,40020252026Release dateArena Text · Korean4o4.1 · Apr5 · Aug5.1 · Nov5.4 · Mar5.5

Benchmarks

in Chat & Knowledge
AA Intelligence Index 4.3.2≈24.7#87
AA Omniscience Index5.4#49
AA Omniscience · Accuracy37.7#54
AA Omniscience · Hallucination Rate51.9#94
Humanity's Last Exam28.5#88
GPQA Diamond87.3#55
AA-LCR (long context) 1.180#42
IFBench (instructions)72.9#34
CritPt (physics)4.9#79
AIME 2025 · AA archived94#9
ARC-AGI-1 · public evaluation77.1#37
ARC-AGI-2 · public evaluation18.3#37
ARC-AGI-1 · semi-private72.8#39
ARC-AGI-2 · semi-private17.6#42
Arena Text · overall1,456#43
Arena Text · overall · full cohort1,442.7#45
Arena Text · Chinese1,495.1#43
Arena Text · English1,450.9#47
Arena Text · French1,450#56
Arena Text · German1,438.1#45
Arena Text · instruction following1,442.6#41
Arena Text · Japanese1,453.2#18
Arena Text · Korean1,400.5#41
Arena Text · non-English1,431.3#44
Arena Text · Polish1,439.7#36
Arena Text · Russian1,435.3#47
Arena Text · Spanish1,432.6#62
MathArena · AIME 202594.2#7
MathArena · Apex1#25
MathArena · Apex Shortlist55.9#21
MathArena · BRUMO 202593.3#7
MathArena · CMIMC 202594.4#3
MathArena · HMMT February 202593.3#5
MathArena · HMMT November 202591.7#5
MMLU-Pro · official boardlab86.4

FAQ

How much does GPT-5.1 cost?›

$1.25 per 1M input tokens and $10 per 1M output tokens; cached input $0.13 per 1M.

Is GPT-5.1 free to use?›

There is no free API; the cheapest paid price is $1.25 per 1M input tokens.

How good is GPT-5.1, and what is it best at?›

#88 of 334 in Chat & Knowledge, #72 of 97 in Coding, #97 of 175 in Agents & Tool Use, #46 of 195 in Writing & Language, #43 of 124 in Vision & Documents. Strongest at: Science, Long context, Knowledge, Instructions.

Is GPT-5.1 better than GPT-5?›

On Arena Text · Korean, GPT-5.1 scores 1,400.5 vs 1,357.4 — higher.

How can I use GPT-5.1?›

Through the API of Azure, OpenAI.

What is the context window of GPT-5.1?›

400K tokens, with up to 128K output tokens.

When was GPT-5.1 released? Is there a newer version?›

Released 2025-11-13. It has been superseded by GPT-6 Sol.

Is GPT-5.1 open source?›

No, GPT-5.1 is a closed model available through an API. The closest open-weights option here is GLM 5.3.

Can GPT-5.1 read images?›

It accepts text and images, PDF files as input.

What is GPT-5.1?›

GPT-5.1 is a closed language model from OpenAI, released Nov 2025, the successor to GPT-5.

What are the best alternatives to GPT-5.1?›

Muse Spark 1.3 (cheaper), GPT-5.6 Luna (better), GLM 5.3 (open weights).