Changelog

Every change we catch, with its date and source.

40 entries · Last check 21 Oct 2026

October 2026

Date	Change	Sources
21 Oct	OpenRouterOffer changed Added the newly listed MiniMax M3.1 Flash free variant under OpenRouter’s shared Free-plan request allowance (20 RPM, 50 RPD by default).	openrouter.aimodels limits
21 Oct	Kilo AI GatewayOffer changed Kilo AI Gateway: added the MiniMax M3.1 Flash free model; Inkling Small’s free variant was withdrawn from the live catalog (only the paid ID remains), so it is now excluded from published offers.	api.kilo.ai kilo.ai
21 Oct	RequestyOffer changed Requesty: added 8 more zero-priced text models (Ling 3.1 Flash, Nemotron 3 Nano Omni, Laguna M.1, Nemotron 3 Ultra, Leanstral 1.5, Gemma 4 31B, Nemotron 3 Super, Nemotron 3 Nano) under the shared 200 requests/day Free plan.	requesty.aipricing ling-3.1-flash nemotron-3-nano-omni-30b-a3b-reasoning laguna-m.1 nemotron-3-ultra-550b-a55b leanstral-1-5 gemma-4-31b-it nemotron-3-super-120b-a12b nemotron-3-nano-30b-a3b
21 Oct	RoutewayOffer changed Routeway: added DeepSeek V4 Flash and Muse Glimmer 30B free variants under the shared Starter 200 requests/day allowance.	routeway.airouteway.ai deepseek-v4-flash:free muse-glimmer-30b:free
21 Oct	SiliconFlowOffer changed SiliconFlow: added 6 more zero-priced text models (Qwen3 8B, Qwen3.5 4B, GLM-4 9B 0414, GLM-Z1 9B 0414, Qwen2.5 7B Instruct, DeepSeek R1 0528 Qwen3 8B) plus DeepSeek-OCR as a specialized, research-only entry.	siliconflow.cn docs.siliconflow.cn
21 Oct	Cloudflare Workers AIOffer changed Cloudflare Workers AI: recorded 5 new specialized models (Clef, Clef Omni, Clef Flash, Clairic, Moondream 3.1) covered by the shared 10,000 Neurons/day allocation. They stay out of text-generation comparisons.	developers.cloudflare.com
21 Oct	NVIDIA NIM / API CatalogOffer changed NVIDIA: recorded 2 new embedding endpoints (Nemotron 3.5 Embed 1B and 8B) under Developer Program prototyping access. They stay out of text-generation comparisons.	integrate.api.nvidia.com docs.api.nvidia.com
21 Oct	IteraComputeOffer added Added IteraCompute to the confirmed text-generation offer catalog with 1 officially documented model deployment. Unknown limits and requirements remain blank; shared pools are not added across models.	iteracompute.com api.iteracompute.com
21 Oct	Hetzner Inference API (Experiments)Offer added Added Hetzner Inference API (Experiments) to the confirmed text-generation offer catalog with 1 officially documented model deployment. Unknown limits and requirements remain blank; shared pools are not added across models.	docs.hetzner.com
21 Oct	Inception (Mercury)Offer added Added Inception (Mercury) to the confirmed text-generation offer catalog with 3 officially documented model deployments. Unknown limits and requirements remain blank; shared pools are not added across models.	docs.inceptionlabs.aiget-started faq rate-limits
21 Oct	Sakura AI EngineOffer added Added Sakura AI Engine to the confirmed text-generation offer catalog with 2 officially documented model deployments. Unknown limits and requirements remain blank; shared pools are not added across models.	ai.sakura.ad.jp manual.sakura.ad.jp
21 Oct	Regolo.aiOffer added Added Regolo.ai to the confirmed text-generation offer catalog with 9 officially documented model deployments. Unknown limits and requirements remain blank; shared pools are not added across models.	regolo.ai api.regolo.ai
21 Oct	Agnes AIOffer added Added Agnes AI to the confirmed text-generation offer catalog with 2 officially documented model deployments. Unknown limits and requirements remain blank; shared pools are not added across models.	agnes-ai.compricing agnes-30-flash
21 Oct	Public AI Inference UtilityOffer added Added Public AI Inference Utility to the confirmed text-generation offer catalog with 9 officially documented model deployments. Unknown limits and requirements remain blank; shared pools are not added across models.	platform.publicai.coplans models
21 Oct	Berget AIOffer added Added Berget AI to the confirmed text-generation offer catalog with 5 officially documented model deployments. Unknown limits and requirements remain blank; shared pools are not added across models.	berget.ai api.berget.ai
21 Oct	Nous PortalOffer added Added Nous Portal to the confirmed text-generation offer catalog with 6 officially documented model deployments. Unknown limits and requirements remain blank; shared pools are not added across models.	portal.nousresearch.com inference-api.nousresearch.com
21 Oct	Zhipu BigModel (CN)Offer added Added Zhipu BigModel (CN) to the confirmed text-generation offer catalog with 5 officially documented model deployments. Unknown limits and requirements remain blank; shared pools are not added across models.	docs.bigmodel.cnpricing openapi.json
21 Oct	Baidu QianfanOffer added Added Baidu Qianfan to the confirmed text-generation offer catalog with 3 officially documented model deployments. Unknown limits and requirements remain blank; shared pools are not added across models.	cloud.baidu.comImi2rpirg rmh4stp0j
21 Oct	Tencent Cloud TokenHubOffer added Added Tencent Cloud TokenHub to the confirmed text-generation offer catalog with 2 officially documented model deployments. Unknown limits and requirements remain blank; shared pools are not added across models.	cloud.tencent.com130053 130051
21 Oct	iFlytek SparkOffer added Added iFlytek Spark to the confirmed text-generation offer catalog with 1 officially documented model deployment. Unknown limits and requirements remain blank; shared pools are not added across models.	xfyun.cn
21 Oct	Sarvam AIOffer added Added Sarvam AI to the confirmed text-generation offer catalog with 2 officially documented model deployments. Unknown limits and requirements remain blank; shared pools are not added across models.	docs.sarvam.airatelimits pricing
21 Oct	Vercel AI GatewayOffer added Added Vercel AI Gateway to the confirmed text-generation offer catalog with 27 officially documented model deployments. Unknown limits and requirements remain blank; shared pools are not added across models.	vercel.compricing models ai-gateway.vercel.sh
21 Oct	Alibaba Cloud Model StudioOffer added Alibaba Cloud Model Studio (Singapore, International scope): published the new-user free quota of 1,000,000 tokens per model, valid 90 days, for 47 verified text-generation model IDs (Qwen, DeepSeek and GLM). Each model has its own pool; billing continues pay-as-you-go after the quota unless Free Quota Only is enabled.	alibabacloud.comnew-free-quota model-pricing
21 Oct	OpenRouterOffer changed Added the newly listed Recursive Flint v1.0 free variant (262,144 context, 235,929 max output) under OpenRouter’s shared Free-plan request allowance (20 RPM, 50 RPD by default).	openrouter.aimodels limits
21 Oct	Google AI Studio / Gemini APIVerification Google Gemini API: re-confirmed 1,048,576 input and 65,536 output token limits on the official model pages for Gemini 3.8 Flash, 3.6 Flash, 3.5 Flash-Lite, 3.1 Flash-Lite, 3 Flash Preview, 2.5 Pro, 2.5 Flash and 2.5 Flash-Lite. Free-tier RPM/TPM/RPD remain unpublished (shown only in AI Studio) and stay unknown.	ai.google.devgemini-3.8-flash gemini-3.6-flash gemini-3.5-flash-lite gemini-3.1-flash-lite gemini-3-flash-preview gemini-2.5-pro gemini-2.5-flash gemini-2.5-flash-lite
21 Oct	Baidu QianfanCorrection Baidu Qianfan: the 1,000,000-token new-user grant is per model, so its rows are no longer labelled as one shared pool.	cloud.baidu.com
17 Oct	Nebius Token FactoryOffer added Added Nebius Token Factory to the confirmed text-generation offer catalog with 1 officially documented model deployment. Unknown limits and requirements remain blank; credits are shared, and trial or upstream restrictions apply as documented.	nebius.com
17 Oct	Vikasit AIOffer added Added Vikasit AI to the confirmed text-generation offer catalog with 1 officially documented model deployment. Unknown limits and requirements remain blank; credits are shared, and trial or upstream restrictions apply as documented.	vikasit.ai
17 Oct	RoutewayOffer added Added Routeway to the confirmed text-generation offer catalog with 1 officially documented model deployment. Unknown limits and requirements remain blank; credits are shared, and trial or upstream restrictions apply as documented.	routeway.airouteway.ai minimax-m2.7:free
17 Oct	PrivatemodeOffer added Added Privatemode to the confirmed text-generation offer catalog with 3 officially documented model deployments. Unknown limits and requirements remain blank; credits are shared, and trial or upstream restrictions apply as documented.	docs.privatemode.ai
17 Oct	UpstageOffer added Added Upstage to the confirmed text-generation offer catalog with 1 officially documented model deployment. Unknown limits and requirements remain blank; credits are shared, and trial or upstream restrictions apply as documented.	console.upstage.ai
17 Oct	NLP CloudOffer added Added NLP Cloud to the confirmed text-generation offer catalog with 6 officially documented model deployments. Unknown limits and requirements remain blank; credits are shared, and trial or upstream restrictions apply as documented.	nlpcloud.com
17 Oct	Kilo AI GatewayOffer added Added Kilo AI Gateway to the confirmed text-generation offer catalog with 14 officially documented model deployments. Unknown limits and requirements remain blank; credits are shared, and trial or upstream restrictions apply as documented.	kilo.ai api.kilo.ai
17 Oct	SiliconFlowOffer added Added SiliconFlow to the confirmed text-generation offer catalog with 2 officially documented model deployments. Unknown limits and requirements remain blank; credits are shared, and trial or upstream restrictions apply as documented.	siliconflow.cn
17 Oct	RequestyOffer added Added Requesty to the confirmed text-generation offer catalog with 3 officially documented model deployments. Unknown limits and requirements remain blank; credits are shared, and trial or upstream restrictions apply as documented.	requesty.aipricing laguna-xs.2 muse-glimmer-30b nemotron-3.5-lightning-30b-a3b
17 Oct	OpenRouterOffer changed Added the newly verified Xing4.0 29B A4B free model variant under OpenRouter’s shared Free-plan request allowance.	openrouter.aimodels pricing how-to-get-the-lowest-cost-llm-inference-on-openrouter
17 Oct	Scaleway Generative APIsOffer changed Confirmed Whisper Large V3 eligibility for the separate free audio-transcription allowance. Speech remains excluded from text-generation comparisons; no cadence is inferred.	scaleway.com
15 Oct	25 providersAudit Initial official-source audit. Retired/paid-only and unconfirmed offers remain visible but are excluded from recurring-free ranking. No historical limits carried forward without confirmation. All 25 providers Groq Cerebras Google AI Studio / Gemini API Mistral AI Studio OpenRouter GitHub Models Cloudflare Workers AI SambaNova Together AI Hugging Face Inference Providers NVIDIA NIM / API Catalog Cohere Chutes Fireworks AI DeepInfra xAI Scaleway Generative APIs OVHcloud AI Endpoints Alibaba Cloud Model Studio Z.ai Nebius Token Factory Novita AI Hyperbolic Replicate AI21 Studio	github.blog inference-docs.cerebras.ai huggingface.co support.together.ai
15 Oct	4 providersVerification Documented Google Pacific-midnight RPD, OpenRouter UTC RPD, Cloudflare UTC Neuron resets, and Cerebras continuous token-bucket replenishment. Added review contracts, sponsored placement separation, static API and quota tracking. All 4 providers Google AI Studio / Gemini API OpenRouter Cloudflare Workers AI Cerebras	ai.google.dev openrouter.ai developers.cloudflare.com inference-docs.cerebras.ai
15 Oct	25 providersVerification Full-catalog re-audit expands all documented free and candidate deployments, records task/eligibility conflicts, confirms Groq Enterprise-only Llama labeling, recovers OpenRouter embedded 20 RPM and conditional daily limits, and introduces explicit monthly /30 normalization. Includes all 19 Scaleway pricing rows as one-time trial candidates, Alibaba console-dependent examples, both AI21 price rows, and NVIDIA catalog/deprecation conflicts. Raw quotas remain unchanged by calculation policy v2.0.0. All 25 providers Groq Cerebras Google AI Studio / Gemini API Mistral AI Studio OpenRouter GitHub Models Cloudflare Workers AI SambaNova Together AI Hugging Face Inference Providers NVIDIA NIM / API Catalog Cohere Chutes Fireworks AI DeepInfra xAI Scaleway Generative APIs OVHcloud AI Endpoints Alibaba Cloud Model Studio Z.ai Nebius Token Factory Novita AI Hyperbolic Replicate AI21 Studio	console.groq.com openrouter.ai ai.google.dev developers.cloudflare.com integrate.api.nvidia.com

Where things stand

Retired (1)

GitHub Models · Fully retired July 30, 2026, including inference API, catalog, playground and BYOK. Historical endpoints and quotas must not be ranked.

No free tier now (2)

Hugging Face Inference Providers · Current free accounts receive no monthly credits. PRO/Team/Enterprise compute credits require a paid subscription; BYOK uses the upstream provider balance.
Together AI · No current free trial. Access requires at least a $5 credit purchase; historical free model IDs are excluded.

Evaluation only (4)

Cohere · Free evaluation API keys have a shared 1,000-call monthly allowance. Production requires a production key; not a renewed daily token quota.
Hetzner Inference API (Experiments) · Experimental: the Inference API is free only while it remains in experimental status; Hetzner promises advance email notice before that changes. Not for production (no performance or availability guarantee). Rate limits are per API key: 10 requests, 4M input tokens and 100k output tokens per 60 seconds. No daily cap is published. Account and payment prerequisites are not stated on the reviewed page.
NVIDIA NIM / API Catalog · Developer Program provides hosted endpoint access for prototyping. No current universal numerical allowance confirmed; do not confuse hosted access with self-hosted NIM licensing.
OVHcloud AI Endpoints · Only Qwen3Guard moderation models are confirmed free during beta. No beta end date, numeric quota or card requirement is established. These specialized models are excluded from the text-generation offer list. A separate first-project cloud-credit promotion is not treated as a recurring allowance.

One-time credits, not daily (14)

AI21 Studio · Official pricing advertises a $10 trial lasting seven days with no card. Current trial model roster and quotas require further verification.
Alibaba Cloud Model Studio · New-user free quota (Singapore region, deployment scope International only): 1,000,000 tokens (input + output combined) for EACH eligible model, auto-granted on first activation, valid 90 days from activation, model release or approval (whichever is later). Pools are independent per model, never summed across models, and shared by the Alibaba Cloud account and its RAM users. Covers real-time inference only (not batch, fine-tuning, deployment or custom models). Catch: after a model's quota is used up, billing continues pay-as-you-go automatically unless Free Quota Only is enabled (it is off by default). Account information must be completed before activation. OAuth has a separate 2,000 calls/day allowance that is not recorded here. Use the Singapore workspace-specific endpoint and substitute workspace_id.
Baidu Qianfan · New-user free token grant (since 2025-10-24, no end date stated): 1,000,000 tokens for each listed model, valid 3 months, online inference only (not batch). Pools are separate per model and must never be summed. Platform default RPM/TPM flow-control values are not free-tier caps and are not recorded as limits.
Berget AI · Trial plan: €5 free starting credits (EUR, so no USD amount is recorded), basic rate limits (not published). A card is required at signup to prevent abuse but nothing is charged. After the credits, an active plan (from €25/month prepaid) is required for continued API access. Expiry is not stated.
Cerebras · Current Free Trial requires a verified payment method: $5, expiring after 30 days. Official FAQ explicitly says there is no permanently free tier.
Inception (Mercury) · One-time signup grant of 100 million tokens shared across all models; no payment details required. Expiry is not stated. Free Tier per-minute limits: 1,000 requests, 1,000,000 input tokens, 100,000 output tokens. Add payment information to continue after the grant runs low.
Nebius Token Factory · Token Factory currently offers initial free credits for API exploration. The official page demonstrates this exact API model within the free onboarding offer. Initial credit amount and expiration are unpublished; card requirement remains unknown.
NLP Cloud · Pay-as-you-go automatically includes a one-time credit for all pre-trained models. Hosted REST API; no OpenAI-compatible base URL is asserted. Expiry and card requirement are unknown.
Public AI Inference Utility · Nonprofit public-AI gateway. Every new account receives $2 of starter credit spent at the listed per-token prices; top up with a payment method to continue. Free tier rate limit: 100 requests/minute per API key. Expiry is not stated. Model contexts are published only as rounded K values.
Regolo.ai · One-time 30-day free trial: 1M tokens per day, stricter (unpublished) rate limits, all Core Models, no credit card. After the trial the account moves to pay-as-you-go with no automatic charge. This is a trial, not a recurring allowance; no total token figure is derived.
Sarvam AI · One-time ₹100 signup credit (INR, so no USD amount is recorded), usable across all Sarvam APIs; credits never expire. Starter-plan rate limits exist but it is not stated that credit-only accounts are on Starter, so RPM is not recorded. Beta models are not included.
Scaleway Generative APIs · New customers receive 1,000,000 tokens once; billing begins after token 1,000,000. Not a daily reset.
Tencent Cloud TokenHub · Free trial package: every language model offers 1,000,000 free tokens, valid 1 year from claim; campaign runs until 2026-12-31. Real-name verification is required. Source conflict: the page says both that all models under one account share the quota and that each model can be claimed once per account, so 1M is treated as an account-level floor and quota scope stays unverified.
Upstage · One-time signup credit, followed by paid API usage; credit expiry and signup payment-card requirements are not established.

Free tier unconfirmed (6)

Chutes · Current page advertises paid subscriptions and per-token billing. Free-trial FAQ answer was not exposed in fetched text; no recurring free allocation confirmed.
DeepInfra · Current pricing is metered. No generally available ongoing free API allowance or signup grant confirmed.
Fireworks AI · Current billing docs require purchasing prepaid credits. No general free inference credit grant confirmed; temporary free training offers are not inference allowances.
Hyperbolic · Current getting-started URL redirects to GPU infrastructure documentation. Previously advertised serverless inference credits and endpoint cannot be reconfirmed.
Replicate · Pay-as-you-go service; a general recurring free LLM API quota is not established. A native predictions API does not establish OpenAI chat compatibility.
xAI · Current API billing is prepaid or invoiced. Historical monthly promotional credits are not treated as a current free offer.

Get alerts

Free AI tokens

October 2026

Date

Change

Sources

21 Oct

OpenRouterOffer changed

Added the newly listed MiniMax M3.1 Flash free variant under OpenRouter’s shared Free-plan request allowance (20 RPM, 50 RPD by default).

openrouter.aimodels limits

21 Oct

Kilo AI GatewayOffer changed

Kilo AI Gateway: added the MiniMax M3.1 Flash free model; Inkling Small’s free variant was withdrawn from the live catalog (only the paid ID remains), so it is now excluded from published offers.

21 Oct

RequestyOffer changed

Requesty: added 8 more zero-priced text models (Ling 3.1 Flash, Nemotron 3 Nano Omni, Laguna M.1, Nemotron 3 Ultra, Leanstral 1.5, Gemma 4 31B, Nemotron 3 Super, Nemotron 3 Nano) under the shared 200 requests/day Free plan.

requesty.aipricing ling-3.1-flash nemotron-3-nano-omni-30b-a3b-reasoning laguna-m.1 nemotron-3-ultra-550b-a55b leanstral-1-5 gemma-4-31b-it nemotron-3-super-120b-a12b nemotron-3-nano-30b-a3b

21 Oct

RoutewayOffer changed

Routeway: added DeepSeek V4 Flash and Muse Glimmer 30B free variants under the shared Starter 200 requests/day allowance.

routeway.airouteway.ai deepseek-v4-flash:free muse-glimmer-30b:free

21 Oct

SiliconFlowOffer changed

SiliconFlow: added 6 more zero-priced text models (Qwen3 8B, Qwen3.5 4B, GLM-4 9B 0414, GLM-Z1 9B 0414, Qwen2.5 7B Instruct, DeepSeek R1 0528 Qwen3 8B) plus DeepSeek-OCR as a specialized, research-only entry.

21 Oct

Cloudflare Workers AIOffer changed

Cloudflare Workers AI: recorded 5 new specialized models (Clef, Clef Omni, Clef Flash, Clairic, Moondream 3.1) covered by the shared 10,000 Neurons/day allocation. They stay out of text-generation comparisons.

developers.cloudflare.com

21 Oct

NVIDIA NIM / API CatalogOffer changed

NVIDIA: recorded 2 new embedding endpoints (Nemotron 3.5 Embed 1B and 8B) under Developer Program prototyping access. They stay out of text-generation comparisons.

21 Oct

IteraComputeOffer added

Added IteraCompute to the confirmed text-generation offer catalog with 1 officially documented model deployment. Unknown limits and requirements remain blank; shared pools are not added across models.

21 Oct

Hetzner Inference API (Experiments)Offer added

Added Hetzner Inference API (Experiments) to the confirmed text-generation offer catalog with 1 officially documented model deployment. Unknown limits and requirements remain blank; shared pools are not added across models.

docs.hetzner.com

21 Oct

Inception (Mercury)Offer added

Added Inception (Mercury) to the confirmed text-generation offer catalog with 3 officially documented model deployments. Unknown limits and requirements remain blank; shared pools are not added across models.

docs.inceptionlabs.aiget-started faq rate-limits

21 Oct

Sakura AI EngineOffer added

Added Sakura AI Engine to the confirmed text-generation offer catalog with 2 officially documented model deployments. Unknown limits and requirements remain blank; shared pools are not added across models.

21 Oct

Regolo.aiOffer added

Added Regolo.ai to the confirmed text-generation offer catalog with 9 officially documented model deployments. Unknown limits and requirements remain blank; shared pools are not added across models.

21 Oct

Agnes AIOffer added

Added Agnes AI to the confirmed text-generation offer catalog with 2 officially documented model deployments. Unknown limits and requirements remain blank; shared pools are not added across models.

agnes-ai.compricing agnes-30-flash

21 Oct

Public AI Inference UtilityOffer added

Added Public AI Inference Utility to the confirmed text-generation offer catalog with 9 officially documented model deployments. Unknown limits and requirements remain blank; shared pools are not added across models.

platform.publicai.coplans models

21 Oct

Berget AIOffer added

Added Berget AI to the confirmed text-generation offer catalog with 5 officially documented model deployments. Unknown limits and requirements remain blank; shared pools are not added across models.

21 Oct

Nous PortalOffer added

Added Nous Portal to the confirmed text-generation offer catalog with 6 officially documented model deployments. Unknown limits and requirements remain blank; shared pools are not added across models.

21 Oct

Zhipu BigModel (CN)Offer added

Added Zhipu BigModel (CN) to the confirmed text-generation offer catalog with 5 officially documented model deployments. Unknown limits and requirements remain blank; shared pools are not added across models.

docs.bigmodel.cnpricing openapi.json

21 Oct

Baidu QianfanOffer added

Added Baidu Qianfan to the confirmed text-generation offer catalog with 3 officially documented model deployments. Unknown limits and requirements remain blank; shared pools are not added across models.

cloud.baidu.comImi2rpirg rmh4stp0j

21 Oct

Tencent Cloud TokenHubOffer added

Added Tencent Cloud TokenHub to the confirmed text-generation offer catalog with 2 officially documented model deployments. Unknown limits and requirements remain blank; shared pools are not added across models.

cloud.tencent.com130053 130051

21 Oct

iFlytek SparkOffer added

Added iFlytek Spark to the confirmed text-generation offer catalog with 1 officially documented model deployment. Unknown limits and requirements remain blank; shared pools are not added across models.

xfyun.cn

21 Oct

Sarvam AIOffer added

Added Sarvam AI to the confirmed text-generation offer catalog with 2 officially documented model deployments. Unknown limits and requirements remain blank; shared pools are not added across models.

docs.sarvam.airatelimits pricing

21 Oct

Vercel AI GatewayOffer added

Added Vercel AI Gateway to the confirmed text-generation offer catalog with 27 officially documented model deployments. Unknown limits and requirements remain blank; shared pools are not added across models.

vercel.compricing models
ai-gateway.vercel.sh

21 Oct

Alibaba Cloud Model StudioOffer added

Alibaba Cloud Model Studio (Singapore, International scope): published the new-user free quota of 1,000,000 tokens per model, valid 90 days, for 47 verified text-generation model IDs (Qwen, DeepSeek and GLM). Each model has its own pool; billing continues pay-as-you-go after the quota unless Free Quota Only is enabled.

alibabacloud.comnew-free-quota model-pricing

21 Oct

OpenRouterOffer changed

Added the newly listed Recursive Flint v1.0 free variant (262,144 context, 235,929 max output) under OpenRouter’s shared Free-plan request allowance (20 RPM, 50 RPD by default).

openrouter.aimodels limits

21 Oct

Google AI Studio / Gemini APIVerification

Google Gemini API: re-confirmed 1,048,576 input and 65,536 output token limits on the official model pages for Gemini 3.8 Flash, 3.6 Flash, 3.5 Flash-Lite, 3.1 Flash-Lite, 3 Flash Preview, 2.5 Pro, 2.5 Flash and 2.5 Flash-Lite. Free-tier RPM/TPM/RPD remain unpublished (shown only in AI Studio) and stay unknown.

ai.google.devgemini-3.8-flash gemini-3.6-flash gemini-3.5-flash-lite gemini-3.1-flash-lite gemini-3-flash-preview gemini-2.5-pro gemini-2.5-flash gemini-2.5-flash-lite

21 Oct

Baidu QianfanCorrection

Baidu Qianfan: the 1,000,000-token new-user grant is per model, so its rows are no longer labelled as one shared pool.

cloud.baidu.com

17 Oct

Nebius Token FactoryOffer added

Added Nebius Token Factory to the confirmed text-generation offer catalog with 1 officially documented model deployment. Unknown limits and requirements remain blank; credits are shared, and trial or upstream restrictions apply as documented.

nebius.com

17 Oct

Vikasit AIOffer added

Added Vikasit AI to the confirmed text-generation offer catalog with 1 officially documented model deployment. Unknown limits and requirements remain blank; credits are shared, and trial or upstream restrictions apply as documented.

vikasit.ai

17 Oct

RoutewayOffer added

Added Routeway to the confirmed text-generation offer catalog with 1 officially documented model deployment. Unknown limits and requirements remain blank; credits are shared, and trial or upstream restrictions apply as documented.

routeway.airouteway.ai minimax-m2.7:free

17 Oct

PrivatemodeOffer added

Added Privatemode to the confirmed text-generation offer catalog with 3 officially documented model deployments. Unknown limits and requirements remain blank; credits are shared, and trial or upstream restrictions apply as documented.

docs.privatemode.ai

17 Oct

UpstageOffer added

Added Upstage to the confirmed text-generation offer catalog with 1 officially documented model deployment. Unknown limits and requirements remain blank; credits are shared, and trial or upstream restrictions apply as documented.

console.upstage.ai

17 Oct

NLP CloudOffer added

Added NLP Cloud to the confirmed text-generation offer catalog with 6 officially documented model deployments. Unknown limits and requirements remain blank; credits are shared, and trial or upstream restrictions apply as documented.

nlpcloud.com

17 Oct

Kilo AI GatewayOffer added

Added Kilo AI Gateway to the confirmed text-generation offer catalog with 14 officially documented model deployments. Unknown limits and requirements remain blank; credits are shared, and trial or upstream restrictions apply as documented.

17 Oct

SiliconFlowOffer added

Added SiliconFlow to the confirmed text-generation offer catalog with 2 officially documented model deployments. Unknown limits and requirements remain blank; credits are shared, and trial or upstream restrictions apply as documented.

siliconflow.cn

17 Oct

RequestyOffer added

Added Requesty to the confirmed text-generation offer catalog with 3 officially documented model deployments. Unknown limits and requirements remain blank; credits are shared, and trial or upstream restrictions apply as documented.

requesty.aipricing laguna-xs.2 muse-glimmer-30b nemotron-3.5-lightning-30b-a3b

17 Oct

OpenRouterOffer changed

Added the newly verified Xing4.0 29B A4B free model variant under OpenRouter’s shared Free-plan request allowance.

openrouter.aimodels pricing how-to-get-the-lowest-cost-llm-inference-on-openrouter

17 Oct

Scaleway Generative APIsOffer changed

Confirmed Whisper Large V3 eligibility for the separate free audio-transcription allowance. Speech remains excluded from text-generation comparisons; no cadence is inferred.

scaleway.com

15 Oct

25 providersAudit

Initial official-source audit. Retired/paid-only and unconfirmed offers remain visible but are excluded from recurring-free ranking. No historical limits carried forward without confirmation.

All 25 providers

Groq Cerebras Google AI Studio / Gemini API Mistral AI Studio OpenRouter GitHub Models Cloudflare Workers AI SambaNova Together AI Hugging Face Inference Providers NVIDIA NIM / API Catalog Cohere Chutes Fireworks AI DeepInfra xAI Scaleway Generative APIs OVHcloud AI Endpoints Alibaba Cloud Model Studio Z.ai Nebius Token Factory Novita AI Hyperbolic Replicate AI21 Studio

15 Oct

4 providersVerification

Documented Google Pacific-midnight RPD, OpenRouter UTC RPD, Cloudflare UTC Neuron resets, and Cerebras continuous token-bucket replenishment. Added review contracts, sponsored placement separation, static API and quota tracking.

All 4 providers

Google AI Studio / Gemini API OpenRouter Cloudflare Workers AI Cerebras

15 Oct

25 providersVerification

Full-catalog re-audit expands all documented free and candidate deployments, records task/eligibility conflicts, confirms Groq Enterprise-only Llama labeling, recovers OpenRouter embedded 20 RPM and conditional daily limits, and introduces explicit monthly /30 normalization. Includes all 19 Scaleway pricing rows as one-time trial candidates, Alibaba console-dependent examples, both AI21 price rows, and NVIDIA catalog/deprecation conflicts. Raw quotas remain unchanged by calculation policy v2.0.0.

All 25 providers