Batch API AI models: about half price if the answer can wait

Prices from official vendor price lists · October 2026
140 models offering asynchronous processing at roughly half the standard rate — the biggest single lever on cost. Output prices run from $0.1 to $180 per 1M tokens across the 132 models with a paid per-token price; 79 have a free tier and 68 ship open weights.
Batch API
Services found: 140
#AI modelCapabilitiesAPI $/1M (in / out)Context windowPrice / moOutput $/1MOverviewVisit site
61

Gemini 3.1 Flash-Lite

Google
Text & chatVisionAudio
$0.25 / $1.51.05M tokensAPI only1.5View
62

Nano Banana 2

Google
VisionImage generationStructured JSON output
API onlyView
63

Nano Banana 2 Lite

Google
VisionImage generationStructured JSON output
API onlyView
64

Nano Banana Pro

Google
VisionImage generationStructured JSON output
$2 / $12API only12View
65

Gemini Omni Flash

Google
Text & chatVisionAudio
$1.5 / $9API only9View
66

Gemini 3.5 Transcribe

Google
AudioAdjustable thinkingSpeech-to-text
API onlyView
67

Gemini 3.5 Transcribe Live

Google
AudioAdjustable thinkingSpeech-to-text
API onlyView
68

Grok 4.20 Reasoning

xAI
Text & chatVisionReasoning
$1.25 / $2.51M tokensAPI only2.5View
69

Grok 4.20 Non-Reasoning

xAI
Text & chatVisionReasoning
$1.25 / $2.51M tokensAPI only2.5View
70

Grok Build 0.1

xAI
Text & chatAdjustable thinkingWeb search grounding
$1 / $2256K tokensAPI only2View
71

DeepSeek V4 Pro

DeepSeek
Text & chatReasoningAdjustable thinking
1M tokensAPI onlyView
72

DeepSeek V4 Flash

DeepSeek
Text & chatVisionReasoning
1M tokensAPI onlyView
73

Codestral

Mistral AI
Text & chatCodingStructured JSON output
$0.3 / $0.9128K tokensAPI only0.9View
74

Devstral 2

Mistral AI
Text & chatCodingStructured JSON output
256K tokensAPI onlyView
75

Magistral Medium 1.2

Mistral AI
Text & chatStructured JSON outputPrompt caching
128K tokensAPI onlyView
76

Qwen3.8 Max

Alibaba
Text & chatReasoningAdjustable thinking
$2 / $61M tokensAPI only6View
77

Qwen3 VL Plus

Alibaba
Text & chatVisionReasoning
$0.2 / $1.6256K tokensAPI only1.6View
78

Qwen3 VL Flash

Alibaba
Text & chatVisionReasoning
$0.05 / $0.4256K tokensAPI only0.4View
79

GLM 5.2

Z.ai
Text & chatReasoningTool calling
$0.966 / $3.0361.05M tokensAPI only3.036View
80

DeepSeek V3.2

DeepSeek
Text & chatStructured JSON outputPrompt caching
$0.269 / $0.4163.8K tokensAPI only0.4View

Browse by need

Reasoning models169 modelsCoding assistants17 modelsVision and images156 modelsLong context 1M+82 modelsFree tier available80 modelsOpen weights88 models

Batch processing queues your requests for asynchronous execution and charges roughly half the standard rate. There is no latency guarantee - results arrive within hours rather than seconds - which makes it the single largest lever on cost for any workload that does not need an immediate answer.

140 models offer it. Typical candidates: overnight document pipelines, bulk classification, dataset labelling, embedding backfills, scheduled reports. Anything driving a live chat window is not a candidate. Combined with prompt caching, a batch pipeline on a mid-tier model lands close to budget-tier pricing while keeping mid-tier quality.

Half price, no promisesRoughly 50% off in exchange for hours of latency instead of seconds.
Stacks with cachingBatch halves everything; caching cuts the repeated input on top of that.
Not for anything liveA user waiting on screen rules batch out entirely.

Models with this capability

The 140 models below all support this capability. Prices are input and output per 1M tokens from the vendor official price list; for dirhams multiply by 3.6725. Tap a card to open the model page.

Claude Fable 5.1
#1
Anthropic
in $10out $501M tokens
GPT-6 Astra
#2
OpenAI
in $10out $501.05M tokens
GPT-6.1 Sol
#3
OpenAI
in $2out $101.05M tokens
GPT-6 Sol
#4
OpenAI
in $2out $101.05M tokens
GPT-6 Luna
#5
OpenAI
in $0.1out $0.51.05M tokens
Claude Opus 5.5
#6
Anthropic
in $4out $201M tokens
Claude Sonnet 5.5
#7
Anthropic
in $2out $101M tokens
Claude Opus 5
#8
Anthropic
in $5out $251M tokens

How to choose

  • Identify which workloads genuinely tolerate delay — most back-office work does.
  • Size batches to the vendor limits rather than sending one giant job.
  • Build retry handling: partial failures inside a batch are normal.
  • Recheck economics after enabling caching; the combination often changes the model choice.

Common mistakes

Batching interactive featuresHalving the cost of something users wait for is a false economy.
Ignoring batch entirelyTeams routinely pay double for pipelines that could run overnight.

Who this suits

Data teamsLabelling, enrichment and backfills over large corpora.
ReportingScheduled summaries and digests that run on a clock.
Migration projectsOne-off passes over an archive where hours are irrelevant.

Related filters

Prompt caching147 models with cheap repeated context
Priority tier65 models with a paid fast lane
Low latency73 small, fast models for volume
Structured JSON151 models with schema-guaranteed output
All 239 modelsThe full rating with filters and prices

Frequently asked questions

How long does a batch take?

Hours rather than seconds. There is no latency guarantee, which is exactly why it costs half.

What workloads suit batch?

Overnight document pipelines, bulk classification, dataset labelling, embedding backfills, scheduled reports. Never a live chat window.

Can I combine batch with caching?

Yes, and it compounds: batch halves the whole bill, caching cuts the repeated input portion by up to 90%.

Is quality identical to standard mode?

Yes — same model, same weights. Only scheduling differs.

Can I cancel a running batch?

Most vendors allow it, though tokens already processed are billed.

Frequently searched

By capabilityReasoningCodingVisionLong context 1M+Free tierOpen weightsLow latencyRealtime voiceImage generation
By vendorAnthropic (13)OpenAI (30)Google DeepMind (17)xAI (8)Meta (6)DeepSeek (9)Mistral AI (11)Alibaba (Qwen) (34)Z.ai (12)Moonshot AI (5)
Popular modelsClaude Fable 5.1GPT-6 AstraClaude Opus 5Claude Fable 5GPT-5.6 SolGLM 5.3Grok 4.6Kimi K3GPT-5.6 Terra
By deliveryBatch APIPrompt cachingPriority tierFrontier tierProprietary APITool callingStructured JSONWeb searchMultilingual

Ratings 0

Avg. score 0

Rate this page

Help us improve