LLMCost
Updated 2026-10-21 · 76 batch SKUs tracked

Batch API pricing: the real discounts

Everyone repeats the same rule of thumb: “batch APIs are half price.” We checked every batch endpoint against its standard twin. Most majors honor it. Several don't — including two providers whose batch costs exactly the same as synchronous.

Honor 50% off
67

SKUs at exactly −50% input

Discount < 50%
2

Cheaper than sync, but stingier

No discount at all
4

Batch = standard price

⚠️ Batch costs the same as standard

Every batch SKU, ranked by real discount

OpenAI: gpt-oss-120bopenai$0.150 / $0.600$0.030 / $0.136−80%
Z.ai: GLM 5.3 Flashz-ai$0.150 / $0.500$0.050 / $0.180−67%
Google: Gemini 3.8 Flashgoogle$0.825 / $4.13$0.375 / $1.88−55%
OpenAI: o4 Miniopenai$1.10 / $4.40$0.550 / $2.20−50%
OpenAI: o3 Miniopenai$1.10 / $4.40$0.550 / $2.20−50%
OpenAI: o3openai$2.00 / $8.00$1.00 / $4.00−50%
OpenAI: GPT-6.1 Sol Proopenai$2.00 / $10.00$1.00 / $5.00−50%
OpenAI: GPT-6.1 Solopenai$2.00 / $10.00$1.00 / $5.00−50%
OpenAI: GPT-6 Sol Proopenai$2.00 / $10.00$1.00 / $5.00−50%
OpenAI: GPT-6 Solopenai$2.00 / $10.00$1.00 / $5.00−50%
OpenAI: GPT-6 Luna Proopenai$0.100 / $0.500$0.050 / $0.250−50%
OpenAI: GPT-6 Lunaopenai$0.100 / $0.500$0.050 / $0.250−50%
OpenAI: GPT-6 Astra Proopenai$10.00 / $50.00$5.00 / $25.00−50%
OpenAI: GPT-6 Astraopenai$10.00 / $50.00$5.00 / $25.00−50%
OpenAI: GPT-5.6 Terra Proopenai$2.00 / $12.00$1.00 / $6.00−50%
OpenAI: GPT-5.6 Terraopenai$2.00 / $12.00$1.00 / $6.00−50%
OpenAI: GPT-5.6 Sol Proopenai$4.00 / $20.00$2.00 / $10.00−50%
OpenAI: GPT-5.6 Solopenai$4.00 / $20.00$2.00 / $10.00−50%
OpenAI: GPT-5.6 Luna Proopenai$0.200 / $1.20$0.100 / $0.600−50%
OpenAI: GPT-5.6 Lunaopenai$0.200 / $1.20$0.100 / $0.600−50%
OpenAI: GPT-5.5 Proopenai$30.00 / $180.00$15.00 / $90.00−50%
OpenAI: GPT-5.5openai$5.00 / $30.00$2.50 / $15.00−50%
OpenAI: GPT-5.4 Proopenai$30.00 / $180.00$15.00 / $90.00−50%
OpenAI: GPT-5.4 Nanoopenai$0.200 / $1.25$0.100 / $0.625−50%
OpenAI: GPT-5.4 Miniopenai$0.750 / $4.50$0.375 / $2.25−50%
OpenAI: GPT-5.4openai$2.50 / $15.00$1.25 / $7.50−50%
OpenAI: GPT-5.2 Proopenai$21.00 / $168.00$10.50 / $84.00−50%
OpenAI: GPT-5.2openai$1.75 / $14.00$0.875 / $7.00−50%
OpenAI: GPT-5.1openai$1.25 / $10.00$0.625 / $5.00−50%
OpenAI: GPT-5 Proopenai$15.00 / $120.00$7.50 / $60.00−50%
OpenAI: GPT-5 Nanoopenai$0.050 / $0.400$0.025 / $0.200−50%
OpenAI: GPT-5 Miniopenai$0.250 / $2.00$0.125 / $1.00−50%
OpenAI: GPT-5openai$1.25 / $10.00$0.625 / $5.00−50%
OpenAI: GPT-4o-miniopenai$0.150 / $0.600$0.075 / $0.300−50%
OpenAI: GPT-4oopenai$2.50 / $10.00$1.25 / $5.00−50%
OpenAI: GPT-4.1 Nanoopenai$0.100 / $0.400$0.050 / $0.200−50%
OpenAI: GPT-4.1 Miniopenai$0.400 / $1.60$0.200 / $0.800−50%
OpenAI: GPT-4.1openai$2.00 / $8.00$1.00 / $4.00−50%
OpenAI: GPT-4 Turboopenai$10.00 / $30.00$5.00 / $15.00−50%
OpenAI: GPT-3.5 Turboopenai$0.500 / $1.50$0.250 / $0.750−50%
Mistral: Mistral Small 4mistralai$0.150 / $0.600$0.075 / $0.300−50%
Mistral: Mistral Medium 3.5mistralai$1.50 / $7.50$0.750 / $3.75−50%
Mistral: Mistral Medium 3.1mistralai$0.400 / $2.00$0.200 / $1.00−50%
Mistral: Mistral Large 3 2512mistralai$0.500 / $1.50$0.250 / $0.750−50%
Mistral: Ministral 3 8B 2512mistralai$0.150 / $0.150$0.075 / $0.075−50%
Mistral: Codestral 2508mistralai$0.300 / $0.900$0.150 / $0.450−50%
Google: Gemini 4 Argongoogle$2.00 / $10.00$1.00 / $5.00−50%
Google: Gemini 3.7 Flashgoogle$0.750 / $3.75$0.375 / $1.88−50%
Google: Gemini 3.6 Flashgoogle$0.750 / $3.75$0.375 / $1.88−50%
Google: Gemini 3.5 Flash Litegoogle$0.300 / $2.50$0.150 / $1.25−50%
Google: Gemini 3.5 Flashgoogle$1.50 / $9.00$0.750 / $4.50−50%
Google: Gemini 3.1 Pro Previewgoogle$2.00 / $12.00$1.00 / $6.00−50%
Google: Gemini 3.1 Flash Litegoogle$0.250 / $1.50$0.125 / $0.750−50%
Google: Gemini 3 Flash Previewgoogle$0.500 / $3.00$0.250 / $1.50−50%
Google: Gemini 2.5 Flash Litegoogle$0.100 / $0.400$0.050 / $0.200−50%
Anthropic: Claude Sonnet 5.5anthropic$2.00 / $10.00$1.00 / $5.00−50%
Anthropic: Claude Sonnet 5anthropic$2.00 / $10.00$1.00 / $5.00−50%
Anthropic: Claude Sonnet 4.6anthropic$3.00 / $15.00$1.50 / $7.50−50%
Anthropic: Claude Sonnet 4.5anthropic$3.00 / $15.00$1.50 / $7.50−50%
Anthropic: Claude Opus 5.5anthropic$4.00 / $20.00$2.00 / $10.00−50%
Anthropic: Claude Opus 5anthropic$5.00 / $25.00$2.50 / $12.50−50%
Anthropic: Claude Opus 4.8anthropic$5.00 / $25.00$2.50 / $12.50−50%
Anthropic: Claude Opus 4.7anthropic$5.00 / $25.00$2.50 / $12.50−50%
Anthropic: Claude Opus 4.6anthropic$5.00 / $25.00$2.50 / $12.50−50%
Anthropic: Claude Opus 4.5anthropic$5.00 / $25.00$2.50 / $12.50−50%
Anthropic: Claude Opus 4.1anthropic$15.00 / $75.00$7.50 / $37.50−50%
Anthropic: Claude Haiku 5.5anthropic$0.100 / $0.500$0.050 / $0.250−50%
Anthropic: Claude Haiku 4.5anthropic$1.00 / $5.00$0.500 / $2.50−50%
Anthropic: Claude Fable 5.1anthropic$10.00 / $50.00$5.00 / $25.00−50%
Anthropic: Claude Fable 5anthropic$10.00 / $50.00$5.00 / $25.00−50%
Qwen: Qwen3.7 Plusqwen$0.320 / $1.28$0.230 / $0.920−28%
xAI: Grok 4.3x-ai$1.25 / $2.50$1.00 / $2.00−20%
DeepSeek: DeepSeek V4.1 Flashdeepseek$0.049 / $0.375$0.050 / $0.150+2%
OpenAI: gpt-oss-20bopenai$0.018 / $0.090$0.024 / $0.112+33%
Z.ai: GLM 5.3z-ai$0.079 / $1.40$0.450 / $2.00+470%
MoonshotAI: Kimi K3moonshotai$0.239 / $11.94$2.28 / $11.40+854%

List prices via OpenRouter, verified 2026-10-21. Batch SLAs (24h window) vary by provider — check before committing large jobs.

Try-before-you-buy: paid models with free twins

These 11 models have a :free variant with identical weights ($0 in/out, rate-limited). Validate your prompt quality on free, then switch the model ID to go production.

FAQ

How much do batch APIs actually save?
The common assumption is 50% off standard pricing. Our live data across all providers shows most major models (OpenAI, Anthropic, Google) do honor exactly 50%, but several providers offer far less — and a few charge identical prices for batch and standard endpoints.
What is a batch API?
Batch endpoints process requests asynchronously, usually within 24 hours, at a discount. You submit many prompts at once and retrieve results later. Ideal for evaluation runs, data labeling, backfills, and any workload that is not latency-sensitive.
Which providers have the worst batch discounts?
This page ranks every batch SKU by its real discount versus the same model's standard price, updated daily. Providers at 0% effectively charge full price for asynchronous processing.
Is batch output quality different from standard?
No — batch uses the same model weights. The tradeoff is purely latency (up to 24h) for price. Results are equivalent for non-realtime workloads.

Related: price changes tracker · models under $1 · monthly cost calculator