Is DeepSeek V4 Flash Vision Exp better than GPT-6 Luna?

On accuracy they are statistically tied: DeepSeek V4 Flash Vision Exp scored 71.2% and GPT-6 Luna 72.5% on the same 80 questions from open datasets. Similar accuracy; GPT-6 Luna costs less per question.

Which is cheaper, DeepSeek V4 Flash Vision Exp or GPT-6 Luna?

GPT-6 Luna is cheaper: $0.00025 per question as billed, against $0.0011. List prices: DeepSeek V4 Flash Vision Exp $0.2156 in / $0.6468 out, GPT-6 Luna $0.1 in / $0.5 out per million tokens.

Which is faster, DeepSeek V4 Flash Vision Exp or GPT-6 Luna?

DeepSeek V4 Flash Vision Exp is faster at 170 output tokens per second, against 137. Time to first token: 0.7 s for DeepSeek V4 Flash Vision Exp, 1.2 s for GPT-6 Luna.

Which has the larger context window, DeepSeek V4 Flash Vision Exp or GPT-6 Luna?

Both accept about the same amount of context: 1M and 1.1M tokens.

Head to head · tested 27 Oct 2026

DeepSeek V4 Flash Vision Exp vs GPT-6 Luna: accuracy, speed and price

Similar accuracy; GPT-6 Luna costs less per question.

DeepSeekDeepSeek V4 Flash Vision Exp71.2% score · #24$0.0011 per question170 tok/s

OpenAIGPT-6 Luna72.5% score · #21$0.00025 per question137 tok/s

Verdict

Built from the measurements below

Accuracy is too close to call: DeepSeek V4 Flash Vision Exp scored 71.2% and GPT-6 Luna 72.5% on the same 80 questions, a gap within what chance alone could produce.
GPT-6 Luna is 4.3× cheaper per question as billed ($0.00025 vs $0.0011), reasoning tokens included.
DeepSeek V4 Flash Vision Exp generates text 1.2× faster: 170 tokens per second against 137.
DeepSeek V4 Flash Vision Exp starts answering sooner: 0.7 s to the first token, against 1.2 s for GPT-6 Luna.

Which should you choose?

Choose DeepSeek V4 Flash Vision Exp for faster streaming, a quicker first token and open weights you can run yourself. Choose GPT-6 Luna for the lower cost per correct answer ($0.00034 against $0.0015) and shorter outputs (about 431 tokens per question against 1,301). The widest subject gap is in medicine: GPT-6 Luna answered 6 of 11 correctly, DeepSeek V4 Flash Vision Exp 4 of 11. With so few questions per subject, treat this as a lead to test on your own prompts, not a result.

Accuracy gap -1.2 points (z = -0.18). A winner is called only when z is at least 1.96, the usual 95% threshold. Speed, latency and price need a 10% margin.

Head to head

Same questions, same prompt, same API

Measure	DeepSeek V4 Flash Vision Exp	GPT-6 Luna	Edge
Verification scoreHigher is better	71.2%	72.5%	Too close to call
Cost per questionLower is better	$0.0011	$0.00025	GPT-6 Luna
Output speedHigher is better	170	137	DeepSeek V4 Flash Vision Exp
Time to first tokenLower is better	0.7 s	1.2 s	DeepSeek V4 Flash Vision Exp
End-to-end response timeLower is better	3.2 s	4.3 s	DeepSeek V4 Flash Vision Exp
Tokens per questionLower is better	1,301	431	GPT-6 Luna
Context windowHigher is better	1M	1.1M	Similar

Side by side

Bars are scaled to the larger of the two values. The winner on each measure is marked in bold.

Every measure

DeepSeek V4 Flash Vision Exp first, GPT-6 Luna second

Verification scorehigher is better

DeepSeek V4 Flash Vision Exp71.2%

GPT-6 Luna72.5%

Cost per questionlower is better

DeepSeek V4 Flash Vision Exp$0.0011

GPT-6 Luna$0.00025

Output speedhigher is better

DeepSeek V4 Flash Vision Exp170

GPT-6 Luna137

Time to first tokenlower is better

DeepSeek V4 Flash Vision Exp0.7 s

GPT-6 Luna1.2 s

End-to-end response timelower is better

DeepSeek V4 Flash Vision Exp3.2 s

GPT-6 Luna4.3 s

Tokens per questionlower is better

DeepSeek V4 Flash Vision Exp1,301

GPT-6 Luna431

Context windowhigher is better

DeepSeek V4 Flash Vision Exp1M

GPT-6 Luna1.1M

Score vs. cost per question

Both models in color, every other model in grey

Results by subject and dataset

Questions answered correctly. Small samples: a gap of one or two questions is noise.

Subject	Questions	DeepSeek V4 Flash Vision Exp	GPT-6 Luna
Science	30	25/30	25/30
Engineering	15	10/15	9/15
Medicine	11	4/11	6/11
Economics	9	8/9	7/9
Law	6	2/6	3/6
Humanities & social science	9	8/9	8/9
SuperGPQA dataset	50	31/50	33/50
MMLU-Pro dataset	30	26/30	25/30

Specifications and pricing

From OpenRouter's catalogue, plus the leaderboard's test dates

Specification	DeepSeek V4 Flash Vision Exp	GPT-6 Luna
Creator	DeepSeek	OpenAI
Released	21 Aug 2026	22 Sept 2026
Weights	Open weights	—
Reasoning	Yes	Yes
Input	text, image	file, image, text
Output	text	text
Context window	1M tokens	1.1M tokens
Knowledge cutoff	—	—
List price in / out per 1M	$0.2156 / $0.6468	$0.1 / $0.5
Blended price per 1M (3:1)	$0.32	$0.20
Cheapest host	DeepInfra · $0.2156 / $0.6468	OpenAI · $0.05 / $0.25
Hosts on OpenRouter	3	3
Last tested	26 Oct 2026	25 Oct 2026

Compare any two models →

DeepSeek V4 Flash Vision Exp vs GPT-6 Luna: frequently asked questions

Is DeepSeek V4 Flash Vision Exp better than GPT-6 Luna?: On accuracy they are statistically tied: DeepSeek V4 Flash Vision Exp scored 71.2% and GPT-6 Luna 72.5% on the same 80 questions from open datasets. Similar accuracy; GPT-6 Luna costs less per question.
Which is cheaper, DeepSeek V4 Flash Vision Exp or GPT-6 Luna?: GPT-6 Luna is cheaper: $0.00025 per question as billed, against $0.0011. List prices: DeepSeek V4 Flash Vision Exp $0.2156 in / $0.6468 out, GPT-6 Luna $0.1 in / $0.5 out per million tokens.
Which is faster, DeepSeek V4 Flash Vision Exp or GPT-6 Luna?: DeepSeek V4 Flash Vision Exp is faster at 170 output tokens per second, against 137. Time to first token: 0.7 s for DeepSeek V4 Flash Vision Exp, 1.2 s for GPT-6 Luna.
Which has the larger context window, DeepSeek V4 Flash Vision Exp or GPT-6 Luna?: Both accept about the same amount of context: 1M and 1.1M tokens.