← All Tools

Best GPT-5 Alternatives

GPT-5 by OpenAI is a flagship model priced at $1.25/10 per 1M tokens (in/out). It's on the expensive side — there are cheaper options with similar quality.

GPT-5

OpenAIFlagship

Input

$1.25/1M

Output

$10/1M

Context

128K

Max Output

16K

Why Switch from GPT-5?

128K context vs 200K+ for newer models
Codex variants better for pure coding

Top Alternatives

#1Mistral Large 3MistralFlagship

38% cheaper, comparable performance.

Input

$2/1M

60% more

Output

$5/1M

50% cheaper

Context

128K

Max Output

16K

MMLU-Pro: 83%(-5.5%)HumanEval: 91%(-4.0%)GPQA:
#2Gemini 3.1 ProGoogleFlagship

Higher benchmark scores, 1M context (8x more), 64K max output.

Input

$2/1M

60% more

Output

$12/1M

20% more

Context

1M

Max Output

64K

MMLU-Pro: 91%(+2.5%)HumanEval: 95%(Same)GPQA: 94.3%(+20.8%)
#3Gemini 3 ProGoogleFlagship

Comparable performance, 1M context (8x more), 66K max output.

Input

$2/1M

60% more

Output

$12/1M

20% more

Context

1M

Max Output

66K

MMLU-Pro: 89.8%(+1.3%)HumanEval: 94%(-1.0%)GPQA: 77%(+3.5%)
#4Claude Opus 4.6AnthropicFlagship

Comparable performance.

Input

$5/1M

300% more

Output

$25/1M

150% more

Context

200K

Max Output

32K

MMLU-Pro: 89.5%(+1.0%)HumanEval: 95%(Same)GPQA: 75.5%(+2.0%)
#5GPT-5.3 CodexOpenAIFlagship

Higher benchmark scores, 66K max output.

Input

$2/1M

60% more

Output

$16/1M

60% more

Context

200K

Max Output

66K

MMLU-Pro: 90%(+1.5%)HumanEval: 96.5%(+1.5%)GPQA: 78%(+4.5%)
#6GPT-5.2 CodexOpenAIFlagship

Comparable performance, 66K max output.

Input

$1.75/1M

40% more

Output

$14/1M

40% more

Context

200K

Max Output

66K

MMLU-Pro: 89%(+0.5%)HumanEval: 95.5%(+0.5%)GPQA: 76%(+2.5%)
#7GLM-5Zhipu AIFlagship

Dramatically cheaper (63% less), 128K max output.

Input

$1/1M

20% cheaper

Output

$3.2/1M

68% cheaper

Context

200K

Max Output

128K

MMLU-Pro: 70.4%(-18.1%)HumanEval: 91%(-4.0%)GPQA: 72%(-1.5%)
#8Gemini 2.5 ProGoogleMid-Tier

Comparable performance, 1M context (8x more), 66K max output.

Input

$1.25/1M

Same price

Output

$10/1M

Same price

Context

1M

Max Output

66K

MMLU-Pro: 87.5%(-1.0%)HumanEval: 93.5%(-1.5%)GPQA: 76%(+2.5%)

Full Comparison Table

ModelInput $/1MOutput $/1MContextMMLU-ProHumanEvalScore
Mistral Large 3Mistral$2.0060% more$5.0050% cheaper128K83%-5.5%91%-4.0%95
Gemini 3.1 ProGoogle$2.0060% more$12.0020% more1M91%+2.5%95%Same90
Gemini 3 ProGoogle$2.0060% more$12.0020% more1M89.8%+1.3%94%-1.0%90
Claude Opus 4.6Anthropic$5.00300% more$25.00150% more200K89.5%+1.0%95%Same85
GPT-5.3 CodexOpenAI$2.0060% more$16.0060% more200K90%+1.5%96.5%+1.5%85
GPT-5.2 CodexOpenAI$1.7540% more$14.0040% more200K89%+0.5%95.5%+0.5%85
GLM-5Zhipu AI$1.0020% cheaper$3.2068% cheaper200K70.4%-18.1%91%-4.0%80
Gemini 2.5 ProGoogle$1.25Same price$10.00Same price1M87.5%-1.0%93.5%-1.5%75
Grok 4xAI$3.00140% more$15.0050% more128K86%-2.5%93%-2.0%74
o4-miniOpenAI$1.1012% cheaper$4.4056% cheaper200K85%-3.5%93.5%-1.5%73
o3OpenAI$0.4068% cheaper$1.6084% cheaper200K87%-1.5%94.5%-0.5%70
Claude Sonnet 4.6Anthropic$3.00140% more$15.0050% more200K86%-2.5%94%-1.0%63
Claude Sonnet 4.5Anthropic$3.00140% more$15.0050% more200K84.5%-4.0%93%-2.0%63
Gemini 3 FlashGoogle$0.5060% cheaper$3.0070% cheaper1M78%-10.5%90%-5.0%63
Claude Haiku 4.5Anthropic$0.8036% cheaper$4.0060% cheaper200K69.4%-19.1%88.1%-6.9%60
MiniMax M2.5MiniMax$0.3076% cheaper$1.2088% cheaper200K82%-6.5%90%-5.0%60
GLM-4.7Zhipu AI$0.6052% cheaper$2.2078% cheaper200K84.3%-4.2%58
Mistral Medium 3Mistral$0.4068% cheaper$2.0080% cheaper128K76%-12.5%87%-8.0%58
GPT-4oOpenAI$2.50100% more$10.00Same price128K80.5%-8.0%91%-4.0%55
Gemini 2.5 FlashGoogle$0.1588% cheaper$0.6094% cheaper1M76%-12.5%89.5%-5.5%53
Llama 4 MaverickMeta$0.3175% cheaper$0.8592% cheaper1M80.5%-8.0%90.2%-4.8%53
DeepSeek R1DeepSeek$0.5556% cheaper$2.1978% cheaper128K84%-4.5%92%-3.0%53
DeepSeek V3DeepSeek$0.1489% cheaper$0.2897% cheaper164K78%-10.5%89%-6.0%43
GPT-4o MiniOpenAI$0.1588% cheaper$0.6094% cheaper128K68%-20.5%87.2%-7.8%40
Llama 4 ScoutMeta$0.1886% cheaper$0.6394% cheaper10M74.2%-14.3%86%-9.0%35

Head-to-Head Comparisons

Alternatives for Other Models