Gemma 4 31B API pricing
Updated
Among the 12 providers we checked (prices read 2026-11-14 to 2026-11-15), the lowest blended standard price for Gemma 4 31B is CoreWeave Forge at USD 0.10 per 1M input tokens and USD 0.34 per 1M output tokens. The median across 12 standard prices is USD 0.14 / USD 0.40. Each price is read from the provider's own pricing page.
Does Google sell Gemma 4 31B directly?
Google publishes no price for Gemma 4 31B on its own API that we could find, so this page compares hosts only.
Gemma 4 31B model facts
Facts from Google's model card for google/gemma-4-31B-it and the Hugging Face model API, read 2026-11-16. BroadNet did not verify them, and hosts may serve the model at a different precision. BroadRouter is BroadNet's site.
| Fact | As stated | Source |
|---|---|---|
| Open weights | Yes. The weights are in a public Hugging Face repository (google/gemma-4-31B-it) with no access gate. | Hugging Face model API, checked 2026-11-16Quote“"private": false, "gated": false, 2 .safetensors files listed” |
| Licence | Apache 2.0 (apache-2.0) | Google model card metadata, checked 2026-11-16Quote“license: apache-2.0” |
| Total parameters | 30.7B | Google model card, checked 2026-11-16Quote“Property | E2B | E4B | 12B Unified | 31B Dense … Total Parameters | 2.3B effective (5.1B with embeddings) | 4.5B effective (8B with embeddings) | 11.95B | 30.7B” |
| Context length | 256K tokens | Google model card, checked 2026-11-16Quote“the medium models support 256K” |
| Input and output | Text and image in; text out | Google model card, checked 2026-11-16Quote“Gemma 4 models are multimodal, handling text and image input (with audio supported on E2B, E4B, and 12B) and generating text output.” |
| Tool calling | Yes | Google model card, checked 2026-11-16Quote“Function Calling – Native support for structured tool use” |
| First published on Hugging Face (not necessarily the release date) | 2026-03-11 | Hugging Face model API, checked 2026-11-16Quote“"createdAt": "2026-03-11T18:22:36.000Z"” |
Gemma 4 31B price per provider
USD per 1M tokens. Standard prices are ranked on a blend of 3 input tokens to 1 output token, cheapest first; the dearest standard price is 5.1× the cheapest. Free rows, the maker's serving version and other service tiers follow, tagged and unranked.
| # | Provider | Input | Cached input | Output | Blended 3:1 | Context | Precision | Source |
|---|---|---|---|---|---|---|---|---|
| 1 | CoreWeave Forge cheapestgoogle/gemma-4-31B-itNotes and source quote“Google Gemma 4 31B $0.10 $0.34 -” | USD 0.10 | not published | USD 0.34 | USD 0.16 | 262,144 | not stated | CoreWeave Forge page checked 2026-11-15 re-checked 2026-11-16 |
| 2 | Featherlessgoogle/gemma-4-31B-itNotes and source quoteDeveloper plan (per-token billing, from 50 USD a month in credits). “PRICING: INPUT $0.12 / CACHED $0.1 / OUTPUT $0.36” | USD 0.12 | USD 0.10 | USD 0.36 | USD 0.18 | 32,768 | FP8 | Featherless page checked 2026-11-15 re-checked 2026-11-16 |
| 3 | Venicegoogle-gemma-4-31b-itNotes and source quote“| Google Gemma 4 31B Instruct | `google-gemma-4-31b-it` | $0.12 | $0.36 | $0.09 | — | Private |” | USD 0.12 | USD 0.09 | USD 0.36 | USD 0.18 | 256,000 | FP4 | Venice page checked 2026-11-15 re-checked 2026-11-16 |
| 4 | Amazon BedrockNotes and source quoteTable headed "Regions: US East (N. Virginia), US East (Ohio) and US West (Oregon)". “Gemma 4 31B $0.14 $0.40” | USD 0.14 | not published | USD 0.40 | USD 0.205 | not stated | not stated | Amazon Bedrock page checked 2026-11-14 re-checked 2026-11-16 |
| 5 | Crusoe Cloudgoogle/gemma-4-31b-itNotes and source quote“Gemma 4 31B text, image $0.14 $0.40 $0.14” | USD 0.14 | USD 0.14 | USD 0.40 | USD 0.205 | 262,144 | not stated | Crusoe Cloud page checked 2026-11-15 re-checked 2026-11-16 |
| 6 | FriendliAIgoogle/gemma-4-31B-itNotes and source quoteContext stated as "262K". No cached input price published. “Input $0.14 / 1M tokens Output $0.4 / 1M tokens” | USD 0.14 | not published | USD 0.40 | USD 0.205 | 262,144 | not stated | FriendliAI page checked 2026-11-15 re-checked 2026-11-16 |
| 7 | GMI Cloudgoogle/gemma-4-31b-itNotes and source quote“Price (per 1M tokens) Input $0.14 Output $0.40 Cache Read $0.10 Cache Write -” | USD 0.14 | USD 0.10 | USD 0.40 | USD 0.205 | 262,144 | FP8 | GMI Cloud page checked 2026-11-15 re-checked 2026-11-16 |
| 8 | Novita AIgoogle/gemma-4-31b-itNotes and source quote“Gemma 4 31B 256K $0.14 /Mt $0.4 /Mt” | USD 0.14 | not published | USD 0.40 | USD 0.205 | 262,144 | not stated | Novita AI page checked 2026-11-14 re-checked 2026-11-16 |
| 9 | ParasailNotes and source quote“Gemma 4 31B $0.15 $0.40 $0.06” | USD 0.15 | USD 0.06 | USD 0.40 | USD 0.2125 | not stated | not stated | Parasail page checked 2026-11-14 re-checked 2026-11-16 |
| 10 | DeepInfragoogle/gemma-4-31B-itNotes and source quote“$0.20 in $0.40 out / 1M tokens” | USD 0.20 | not published | USD 0.40 | USD 0.25 | 262,144 | FP8 | DeepInfra page checked 2026-11-14 re-checked 2026-11-16 |
| 11 | SambaNovagemma-4-31B-itNotes and source quote“gemma-4-31B-it N/A 0.38 USD 1.15 USD” | USD 0.38 | not published | USD 1.15 | USD 0.5725 | not stated | not stated | SambaNova page checked 2026-11-14 re-checked 2026-11-16 |
| 12 | SiliconFlowNotes and source quoteContext shown as 262K. “gemma-4-31B-it 262K $0.75 $0.25 $1.0” | USD 0.75 | USD 0.25 | USD 1 | USD 0.8125 | 262,144 | not stated | SiliconFlow page checked 2026-11-14 re-checked 2026-11-16 |
| — | DeepInfra flex / spotgoogle/gemma-4-31B-itNotes and source quoteFlex service tier (0.8× standard). “Flex (0.8×) $0.16 $0.32” | USD 0.16 | not published | USD 0.32 | USD 0.20 | 262,144 | FP8 | DeepInfra page checked 2026-11-14 re-checked 2026-11-16 |
| — | DeepInfra prioritygoogle/gemma-4-31B-itNotes and source quotePriority service tier (1.5× standard). “Priority (1.5×) $0.30 $0.60” | USD 0.30 | not published | USD 0.60 | USD 0.375 | 262,144 | FP8 | DeepInfra page checked 2026-11-14 re-checked 2026-11-16 |
- Ranked: standard-tier prices only, on (3 × input + output) ÷ 4. "Cached input: not published" means the provider's page shows no cache price; budget cached tokens at the input price.
- Other tiers (flex / spot and priority): batch runs asynchronously within hours; flex and spot are best-effort and can be slow or refused at capacity; priority and fast cost more for reliability or speed. Shown, not ranked.
- Automated re-check 2026-11-16: a weekly job re-fetches each source page and looks for the recorded quote or prices. It is not a human review. "Re-checked" means it still found them. Rows with no re-check note show only the date the price was read.
- Precision is shown only where the provider states it; "not stated" is unknown, not full precision.
Report a wrong price or number (email to info@broadnet.ai, with this page's link filled in).
What does Gemma 4 31B cost for a real workload?
Monthly cost over 30 days at each standard price, using the presets of the LLM API cost calculator. Providers without a cached price bill cached input at the input price.
- Support chatbot: 2,000 requests a day, 1,500 input tokens (60% cached), 300 output.
- RAG over documents: 1,000 requests a day, 8,000 input tokens (30% cached), 500 output.
- Coding agent (one developer): 1,500 requests a day, 40,000 input tokens (85% cached), 2,000 output incl. reasoning.
| Provider | Support chatbot | RAG over documents | Coding agent (one developer) |
|---|---|---|---|
| CoreWeave Forge | USD 15.12 | USD 29.10 | USD 210.60 |
| Featherless | USD 16.20 | USD 32.76 | exceeds context |
| Venice | USD 15.66 | USD 32.04 | USD 202.50 |
| Amazon Bedrock | USD 19.80 | USD 39.60 | USD 288.00 |
| Crusoe Cloud | USD 19.80 | USD 39.60 | USD 288.00 |
| FriendliAI | USD 19.80 | USD 39.60 | USD 288.00 |
| GMI Cloud | USD 17.64 | USD 36.72 | USD 226.80 |
| Novita AI | USD 19.80 | USD 39.60 | USD 288.00 |
| Parasail | USD 15.84 | USD 35.52 | USD 168.30 |
| DeepInfra | USD 25.20 | USD 54.00 | USD 396.00 |
| SambaNova | USD 54.90 | USD 108.45 | USD 787.50 |
| SiliconFlow | USD 58.50 | USD 159.00 | USD 675.00 |
For the support chatbot, the cheapest price (CoreWeave Forge, USD 15.12 a month) saves USD 4.68 against the median (USD 19.80). For the coding agent the gap is USD 119.70 a month (42% lower). The winner changes with the workload: Parasail wins the cache-heavy agent, CoreWeave Forge the chatbot.
Which Gemma 4 31B provider is cheapest for your traffic?
CoreWeave Forge has the lowest input price (USD 0.10) and the lowest output price (USD 0.34), so it is the cheapest standard price at every mix of input and output tokens, before cache discounts.
| Provider | No cache hits | 50% cached | 90% cached |
|---|---|---|---|
| CoreWeave Forge (no cache price) | USD 0.16 cheapest | USD 0.16 cheapest | USD 0.16 |
| Featherless | USD 0.18 | USD 0.1725 | USD 0.1665 |
| Venice | USD 0.18 | USD 0.1688 | USD 0.1598 |
| Amazon Bedrock (no cache price) | USD 0.205 | USD 0.205 | USD 0.205 |
| Crusoe Cloud | USD 0.205 | USD 0.205 | USD 0.205 |
| FriendliAI (no cache price) | USD 0.205 | USD 0.205 | USD 0.205 |
| Parasail | USD 0.2125 | USD 0.1788 | USD 0.1517 cheapest |
USD per 1M tokens, 3 input tokens to 1 output token, each provider's cheapest standard row; 5 dearer providers are in the price table above. A provider without a cached price bills cached input at the input price. Any charge for writing to the cache, which some providers add, is not included.
5 of 12 providers publish a cached-input price below the input price: Featherless (17% off), Venice (25% off), GMI Cloud (29% off), Parasail (60% off), SiliconFlow (67% off). CoreWeave Forge, Amazon Bedrock, FriendliAI, Novita AI, DeepInfra and SambaNova publish none, so cached tokens there bill at the input price. Crusoe Cloud lists a cached price equal to its input price.
Context windows differ: 262,144 tokens at CoreWeave Forge, Crusoe Cloud, FriendliAI and 4 more; 256,000 tokens at Venice; 32,768 tokens at Featherless. Amazon Bedrock, Parasail and SambaNova do not state one. The cheapest row is not always the one that fits a long prompt.
What does Gemma 4 31B cost in other languages?
The same text needs a different number of tokens in each language. With the Gemma 4 VL 26B tokenizer, sending 1M characters costs this much at the cheapest standard input price we found for Gemma 4 31B:
| Content | Input per 1M characters | Cheapest provider |
|---|---|---|
| English | USD 0.022 | CoreWeave Forge |
| Arabic MSA | USD 0.034 | CoreWeave Forge |
| Hindi | USD 0.027 | CoreWeave Forge |
| Chinese | USD 0.067 | CoreWeave Forge |
| Code | USD 0.029 | CoreWeave Forge |
From BroadRouter's real cost by language dataset, which covers 15 languages and content types for every model we could match to a measured tokenizer.
Published benchmark results
BroadNet did not run these benchmarks; results are as published by the Google model card and Epoch AI and may use different settings. BroadRouter is BroadNet's site.
| Benchmark | Result | Setting or note | Source |
|---|---|---|---|
| MMLU Pro | 85.2% | Instruction-tuned model | Google model card, checked 2026-11-16 |
| GPQA Diamond | 84.3% | Instruction-tuned model | Google model card, checked 2026-11-16 |
| LiveCodeBench v6 | 80.0% | Instruction-tuned model | Google model card, checked 2026-11-16 |
| AIME 2026 no tools | 89.2% | Instruction-tuned model | Google model card, checked 2026-11-16 |
| MMMU Pro | 76.9% | Instruction-tuned model | Google model card, checked 2026-11-16 |
| GPQA Diamond | 0.758 | Minimal reasoning setting; Epoch run gemma-4-31b-it_minimal, mean score on a 0 to 1 scale | Epoch AI, checked 2026-11-16 |
Epoch AI rows: Epoch AI, Capabilities & benchmarking, licensed CC BY 4.0, shown as the mean score on a 0 to 1 scale, rounded to three decimals (hover a value for the exact figure in Epoch's data file).
Estimate your own bill
- LLM API cost calculator with Gemma 4 31B preloaded: your traffic and cache rate, against any of 428 models.
- Gemma token counter: count the tokens in your real prompts. Tokenizers differ, so the same text costs a different number of tokens on different models.
- Prompt caching calculator: what a higher cache-hit rate saves.
How does Gemma 4 31B compare with other Google models?
The Google models whose cheapest standard price is closest to Gemma 4 31B's (USD 0.16 blended):
| Model | Cheapest input | Cheapest output | Max context | Providers |
|---|---|---|---|---|
| gemma-4-31B-it-turbo | USD 0.09 | USD 0.34 | 262,144 | 2 |
| Gemini 2.5 Flash-Lite | USD 0.10 | USD 0.40 | — | 2 |
| Gemma 4 26B A4B | USD 0.07 | USD 0.34 | 262,144 | 11 |
| Gemma 3 27B | USD 0.08 | USD 0.16 | 198,000 | 6 |
| Gemma 3 12B | USD 0.05 | USD 0.15 | 131,072 | 4 |
| Gemma 3 4B | USD 0.04 | USD 0.08 | 131,072 | 3 |
Priced closest from other makers: Llama 3.2 11B Vision Instruct (Meta, USD 0.16) · NVIDIA Nemotron 3 Super 120B A12B (NVIDIA, USD 0.1563) · Llama 3.3 70B (Meta, USD 0.155).
Compare Gemma 4 31B with
- Gemma 4 31B vs Gemma 3 27B: API price side by side
- Llama 4 Scout vs Gemma 4 31B: API price side by side
- Gemma 4 26B A4B vs Gemma 4 31B: API price side by side
- Gemma 4 31B vs Qwen3.8 27B: API price side by side
- DeepSeek V4.1 Flash vs Gemma 4 31B: API price side by side
Also compared: vs Gemma 3 27B · vs Llama 4 Scout 17B 16E Instruct · vs Gemma 4 26B A4B · vs Qwen3.8 27B · vs DeepSeek V4.1 Flash.
Every Gemma 4 31B price comparison · All model-vs-model price comparisons
All 428 models: API prices compared · Cheapest LLM APIs right now · LLM API price index (open data) · Download all prices (CSV, JSON) · Price change log
Questions
How much does the Gemma 4 31B API cost per token?
From USD 0.10 per 1M input tokens (USD 0.0000001 per token) and USD 0.34 per 1M output tokens (USD 0.00000034 per token) at the lowest standard prices we found. The median of 12 standard prices is USD 0.14 input and USD 0.40 output. Each price was read from the provider's own pricing page; the source is linked in its row.
Which provider is cheapest for Gemma 4 31B?
Among the 12 providers we checked (prices read 2026-11-14 to 2026-11-15), the lowest is CoreWeave Forge, ranked on a blend of 3 input tokens to 1 output token (USD 0.16 per 1M). It also has the lowest input and the lowest output price. Batch, flex, priority and fast tiers are shown in the table but not ranked.
Does Google sell Gemma 4 31B on its own API?
Google publishes no price for Gemma 4 31B on its own API that we could find, so there is no maker price to compare. The lowest-priced host we checked is CoreWeave Forge at USD 0.10 / USD 0.34 per 1M tokens.
Does Gemma 4 31B have cached input pricing?
6 of 12 standard prices include one. The lowest is USD 0.06 per 1M at Parasail; the median is USD 0.10. Providers without one bill cached input at the input price. What cached input means.
What is the Gemma 4 31B context window?
Up to 262,144 tokens. 2 standard prices come with less, down to 32,768 tokens (Featherless and Venice). Check the context column before you pick the cheapest row for long prompts. Context window calculator.
How much does a coding agent cost per month on Gemma 4 31B?
For 1,500 requests a day, 40,000 input tokens (85% cached), 2,000 output incl. reasoning, over 30 days: USD 168.30 a month at Parasail, USD 288.00 at the median price and USD 787.50 at SambaNova. Try your own numbers.
Is there a batch, flex or priority price for Gemma 4 31B?
Yes: flex / spot from DeepInfra at USD 0.16 / USD 0.32; priority from DeepInfra at USD 0.30 / USD 0.60 per 1M input / output tokens. These tiers trade speed or availability for price (batch runs within hours, flex and spot can be slow or refused at capacity, priority and fast cost more for speed or reliability), so they are shown but not ranked.
Is Gemma 4 31B free to use through an API?
No provider in our data lists a free price for Gemma 4 31B as of 2026-11-15. Free tiers and credits by provider (checked 2026-11-14) lists general offers that do not name a model; all free LLM API models are on one page.
Is Gemma 4 31B open source, with open weights, and what is its licence?
Google publishes the weights on Hugging Face as google/gemma-4-31B-it, in a public repository with no access gate. The licence, as the Google model card metadata states it: Apache 2.0 (apache-2.0). Open weights means you can download and run the model yourself; what you may do with it, including commercial use, is set by that licence, so read it in full (checked 2026-11-16). These facts come from Google's card and the Hugging Face model API; BroadNet did not verify them, and hosts may serve the model at a different precision.
How many parameters does Gemma 4 31B have?
Total parameters, as the Google model card states them: 30.7B. (checked 2026-11-16; BroadNet did not verify these figures.)
Does Gemma 4 31B support tool calling?
Yes, according to the Google model card: “Function Calling – Native support for structured tool use” (checked 2026-11-16). Whether the tools parameter works through an API also depends on the provider serving it, so check that provider's docs. BroadNet did not test it.
How do I call Gemma 4 31B with the OpenAI SDK?
BroadNet does not serve Gemma 4 31B. Use the OpenAI-compatible base URL of a provider that lists it: create the client with OpenAI(base_url=..., api_key=...) and pass that provider's model ID. Documented base URLs for providers on this page, in alphabetical order: DeepInfra (https://api.deepinfra.com/v1/openai, model google/gemma-4-31B-it). Each provider page quotes the provider's docs with a check date.
How does Gemma 4 31B score on benchmarks?
As published, the Google model card reports 85.2% on MMLU Pro and 84.3% on GPQA Diamond; Epoch AI measured 0.758 on GPQA Diamond in its own run, as a mean score on a 0 to 1 scale (checked 2026-11-16). BroadNet did not run these benchmarks, and sources may use different settings, so the figures are not directly comparable. Every published result we found, with its setting and source.
Read as MarkdownAsk ChatGPTAsk ClaudeAsk Perplexity
Sources
- Amazon Bedrock: pricing (checked 2026-11-14)
- CoreWeave Forge: pricing (checked 2026-11-15)
- Crusoe Cloud: pricing (checked 2026-11-15)
- DeepInfra: pricing (checked 2026-11-14)
- Featherless: pricing (checked 2026-11-15)
- FriendliAI: pricing (checked 2026-11-15)
- GMI Cloud: pricing (checked 2026-11-15)
- Novita AI: pricing (checked 2026-11-14)
- Parasail: pricing (checked 2026-11-14)
- SambaNova: pricing (checked 2026-11-14)
- SiliconFlow: pricing (checked 2026-11-14)
- Venice: pricing (checked 2026-11-15)