# The Token Bill

> Live leaderboard of AI API deals, pricing, and SWE-bench scores. Compare the cheapest Claude, GPT, Gemini, Kimi, DeepSeek, Llama and Qwen hosts across 16+ providers. Refreshed daily.

Last generated: 2026-09-02T01:06:43.767Z
Models tracked: 14
Hosts tracked: 16

## Core content

- [Homepage (interactive leaderboard)](https://thetokenbill.com/)
- [Full leaderboard in Markdown](https://thetokenbill.com/index.md)
- [Cheapest AI APIs](https://thetokenbill.com/cheapest)
- [Best open-source models for coding](https://thetokenbill.com/best/open-source-models-for-coding)
- [Best coding models for design](https://thetokenbill.com/best/coding-models-for-design)
- [Best video-understanding models](https://thetokenbill.com/best/video-understanding-models)
- [Alternatives to Kimi K3](https://thetokenbill.com/alternatives/kimi-k3)
- [AI Deals (subscriptions, flat-rate plans, harnesses)](https://thetokenbill.com/ai-deals)
- [About](https://thetokenbill.com/about)

## Provider profiles

- [AKI.IO provider profile](https://thetokenbill.com/providers/aki-io)
- [Alibaba Cloud Model Studio provider profile](https://thetokenbill.com/providers/alibaba-model-studio)
- [Google Vertex AI provider profile](https://thetokenbill.com/providers/google-vertex)
- [GreenPT provider profile](https://thetokenbill.com/providers/greenpt)
- [Inceptron provider profile](https://thetokenbill.com/providers/inceptron)
- [Infercom provider profile](https://thetokenbill.com/providers/infercom)
- [IONOS AI Model Hub provider profile](https://thetokenbill.com/providers/ionos)
- [Lyceum Technology provider profile](https://thetokenbill.com/providers/lyceum)
- [Microsoft Foundry provider profile](https://thetokenbill.com/providers/azure-foundry)
- [Mistral API provider profile](https://thetokenbill.com/providers/mistral-api)
- [Nebius Token Factory provider profile](https://thetokenbill.com/providers/nebius)
- [Nebul provider profile](https://thetokenbill.com/providers/nebul)
- [Regolo provider profile](https://thetokenbill.com/providers/regolo)
- [Scaleway provider profile](https://thetokenbill.com/providers/scaleway)
- [TensorX provider profile](https://thetokenbill.com/providers/tensorx)


## Geographic coverage

- [AI infrastructure in Brazil](https://thetokenbill.com/countries/br)
- [AI infrastructure in Chile](https://thetokenbill.com/countries/cl)
- [AI infrastructure in China](https://thetokenbill.com/countries/cn)
- [AI infrastructure in European Union](https://thetokenbill.com/regions/eu)
- [AI infrastructure in Finland](https://thetokenbill.com/countries/fi)
- [AI infrastructure in France](https://thetokenbill.com/countries/fr)
- [AI infrastructure in Germany](https://thetokenbill.com/countries/de)
- [AI infrastructure in Italy](https://thetokenbill.com/countries/it)
- [AI infrastructure in Singapore](https://thetokenbill.com/countries/sg)
- [AI infrastructure in United Arab Emirates](https://thetokenbill.com/countries/ae)
- [AI infrastructure in United Kingdom](https://thetokenbill.com/countries/gb)
- [AI infrastructure in United States](https://thetokenbill.com/regions/united-states)


## Geographic answer pages

- [AI providers in Brazil](https://thetokenbill.com/countries/br/providers)
- [AI APIs with documented data residency in China](https://thetokenbill.com/countries/cn/data-residency)
- [AI models available in European Union](https://thetokenbill.com/regions/eu/models)
- [AI providers in European Union](https://thetokenbill.com/regions/eu/providers)
- [Cheapest documented AI APIs in European Union](https://thetokenbill.com/regions/eu/cheapest)
- [Largest non-major AI provider in European Union](https://thetokenbill.com/regions/eu/providers/largest)
- [AI APIs with documented data residency in European Union](https://thetokenbill.com/regions/eu/data-residency)
- [Providers for kimi-k3 in European Union](https://thetokenbill.com/regions/eu/models/kimi-k3)
- [Providers for glm-5-2 in European Union](https://thetokenbill.com/regions/eu/models/glm-5-2)
- [Providers for gpt-oss-120b in European Union](https://thetokenbill.com/regions/eu/models/gpt-oss-120b)
- [Providers for mistral-small-3-2-24b-instruct-2506 in European Union](https://thetokenbill.com/regions/eu/models/mistral-small-3-2-24b-instruct-2506)
- [Providers for deepseek-v4-flash-0731 in European Union](https://thetokenbill.com/regions/eu/models/deepseek-v4-flash-0731)
- [Providers for qwen3-8-27b in European Union](https://thetokenbill.com/regions/eu/models/qwen3-8-27b)
- [Providers for minimax-m3 in European Union](https://thetokenbill.com/regions/eu/models/minimax-m3)
- [Providers for kimi-k2-7-code in European Union](https://thetokenbill.com/regions/eu/models/kimi-k2-7-code)
- [Providers for qwen3-5-9b in European Union](https://thetokenbill.com/regions/eu/models/qwen3-5-9b)
- [Providers for qwen3-embedding-8b in European Union](https://thetokenbill.com/regions/eu/models/qwen3-embedding-8b)
- [Providers for apertus-70b in European Union](https://thetokenbill.com/regions/eu/models/apertus-70b)
- [Providers for llama-3-3-70b-instruct in European Union](https://thetokenbill.com/regions/eu/models/llama-3-3-70b-instruct)
- [AI providers in Finland](https://thetokenbill.com/countries/fi/providers)
- [AI providers in France](https://thetokenbill.com/countries/fr/providers)
- [AI providers in Germany](https://thetokenbill.com/countries/de/providers)
- [AI APIs with documented data residency in Germany](https://thetokenbill.com/countries/de/data-residency)
- [AI providers in Singapore](https://thetokenbill.com/countries/sg/providers)


Per-model landing pages: `/{modifier}/{model}` where modifier ∈ {`cheapest`, `best`, `most-reliable`}. Examples:
- https://thetokenbill.com/cheapest/claude-sonnet-4-6
- https://thetokenbill.com/cheapest/kimi-k2-5
- https://thetokenbill.com/cheapest/llama-3-3-70b-instruct
- https://thetokenbill.com/best/gpt-5-4
- https://thetokenbill.com/most-reliable/deepseek-v3-2

## Top 15 cheapest models (by lowest input $/MTok)

1. **GPT-5.6 Luna** — $0.20/MTok input at **OpenAI** · 1 host · 75.7% SWE-bench · [https://thetokenbill.com/cheapest/gpt-5-6-luna](https://thetokenbill.com/cheapest/gpt-5-6-luna)
2. **Gemini 3 Flash Preview** — $0.50/MTok input at **Google AI Studio** · 1 host · 75.4% SWE-bench · [https://thetokenbill.com/cheapest/gemini-3-flash-preview](https://thetokenbill.com/cheapest/gemini-3-flash-preview)
3. **Kimi K2.5** — $0.50/MTok input at **TensorX** · 7 hosts · 73.8% SWE-bench · [https://thetokenbill.com/cheapest/kimi-k2-5](https://thetokenbill.com/cheapest/kimi-k2-5)
4. **Kimi K2.6** — $0.73/MTok input at **Inceptron** · 3 hosts · 76.7% SWE-bench · [https://thetokenbill.com/cheapest/kimi-k2-6](https://thetokenbill.com/cheapest/kimi-k2-6)
5. **GLM 5.2** — $1.20/MTok input at **Inceptron** · 4 hosts · 78.7% SWE-bench · [https://thetokenbill.com/cheapest/glm-5-2](https://thetokenbill.com/cheapest/glm-5-2)
6. **GLM 5** — $1.20/MTok input at **Amazon Bedrock** · 6 hosts · 72.1% SWE-bench · [https://thetokenbill.com/cheapest/glm-5](https://thetokenbill.com/cheapest/glm-5)
7. **Gemini 2.5 Pro** — $1.25/MTok input at **Google AI Studio** · 1 host · 57.6% SWE-bench · [https://thetokenbill.com/cheapest/gemini-2-5-pro](https://thetokenbill.com/cheapest/gemini-2-5-pro)
8. **DeepSeek V4 Pro 0423** — $1.32/MTok input at **Fireworks AI** · 5 hosts · 77.6% SWE-bench · [https://thetokenbill.com/cheapest/deepseek-v4-pro](https://thetokenbill.com/cheapest/deepseek-v4-pro)
9. **GLM 5.1** — $1.40/MTok input at **Fireworks AI** · 1 host · 74.2% SWE-bench · [https://thetokenbill.com/cheapest/glm-5-1](https://thetokenbill.com/cheapest/glm-5-1)
10. **GPT-5.6 Terra** — $2.00/MTok input at **OpenAI** · 1 host · 78.4% SWE-bench · [https://thetokenbill.com/cheapest/gpt-5-6-terra](https://thetokenbill.com/cheapest/gpt-5-6-terra)
11. **Claude Sonnet 4.6** — $3.00/MTok input at **Anthropic** · 1 host · 75.2% SWE-bench · [https://thetokenbill.com/cheapest/claude-sonnet-4-6](https://thetokenbill.com/cheapest/claude-sonnet-4-6)
12. **GPT-5.5** — $5.00/MTok input at **OpenAI** · 1 host · 83.1% SWE-bench · [https://thetokenbill.com/cheapest/gpt-5-5](https://thetokenbill.com/cheapest/gpt-5-5)
13. **Claude Opus 4.7** — $5.00/MTok input at **Anthropic** · 1 host · 83.5% SWE-bench · [https://thetokenbill.com/cheapest/claude-opus-4-7](https://thetokenbill.com/cheapest/claude-opus-4-7)
14. **Claude Opus 4.6** — $5.00/MTok input at **Anthropic** · 1 host · 78.7% SWE-bench · [https://thetokenbill.com/cheapest/claude-opus-4-6](https://thetokenbill.com/cheapest/claude-opus-4-6)

## Top 10 best value models (SWE-bench per dollar)

1. **GPT-5.6 Luna** — 75.7% SWE-bench at $0.37/MTok blended via **OpenAI** · [https://thetokenbill.com/best/gpt-5-6-luna](https://thetokenbill.com/best/gpt-5-6-luna)
2. **Kimi K2.5** — 73.8% SWE-bench at $0.88/MTok blended via **TensorX** · [https://thetokenbill.com/best/kimi-k2-5](https://thetokenbill.com/best/kimi-k2-5)
3. **Gemini 3 Flash Preview** — 75.4% SWE-bench at $0.92/MTok blended via **Google AI Studio** · [https://thetokenbill.com/best/gemini-3-flash-preview](https://thetokenbill.com/best/gemini-3-flash-preview)
4. **Kimi K2.6** — 76.7% SWE-bench at $1.19/MTok blended via **Inceptron** · [https://thetokenbill.com/best/kimi-k2-6](https://thetokenbill.com/best/kimi-k2-6)
5. **GLM 5.2** — 78.7% SWE-bench at $1.70/MTok blended via **Inceptron** · [https://thetokenbill.com/best/glm-5-2](https://thetokenbill.com/best/glm-5-2)
6. **DeepSeek V4 Pro 0423** — 77.6% SWE-bench at $2.03/MTok blended via **Fireworks AI** · [https://thetokenbill.com/best/deepseek-v4-pro](https://thetokenbill.com/best/deepseek-v4-pro)
7. **GLM 5** — 72.1% SWE-bench at $1.64/MTok blended via **Amazon Bedrock** · [https://thetokenbill.com/best/glm-5](https://thetokenbill.com/best/glm-5)
8. **GLM 5.1** — 74.2% SWE-bench at $1.90/MTok blended via **Fireworks AI** · [https://thetokenbill.com/best/glm-5-1](https://thetokenbill.com/best/glm-5-1)
9. **GPT-5.6 Terra** — 78.4% SWE-bench at $3.67/MTok blended via **OpenAI** · [https://thetokenbill.com/best/gpt-5-6-terra](https://thetokenbill.com/best/gpt-5-6-terra)
10. **Gemini 2.5 Pro** — 57.6% SWE-bench at $2.71/MTok blended via **Google AI Studio** · [https://thetokenbill.com/best/gemini-2-5-pro](https://thetokenbill.com/best/gemini-2-5-pro)

## Curated subscription deals

- **Fireworks Fire Pass** — $7/wk · Kimi K2.5 Turbo · BYOH. Flat weekly rate for high-volume Kimi K2.5 Turbo access. API key works in any OpenAI-compatible harness.
- **Synthetic.new — 1 Pack** — $30/mo · Claude-class frontier · BYOH. Flat monthly rate giving API access to Claude-class frontier models. Drop the key into any agent framework.
- **MiniMax Token Plan — Starter** — $10/mo · MiniMax M2.7 · BYOH. Entry tier of MiniMax's flat-rate API bundle. Uses standard API keys so any harness works.
- **MiniMax Token Plan — Plus** — $20/mo · MiniMax M2.7 + speech/image · BYOH. Mid-tier MiniMax subscription with 4,500 requests per 5-hour window. Best price/volume ratio in the flat-rate bucket.
- **MiniMax Token Plan — Max** — $50/mo · MiniMax M2.7 + all modalities · BYOH. Top MiniMax tier, 15,000 requests per 5-hour window plus all modalities. For continuous agent workflows.
- **OpenCode Zen** — PAYG ~$20 topup · Kimi K2.5, GPT-5-Codex, Gemini, Claude · BYOH. Pay-as-you-go gateway to multiple frontier models at zero markup. No monthly fee — just top up and use.
- **Claude Pro** — $20/mo · Claude Sonnet 4.6 · BYOH. Standard Claude chat subscription. Claude Code subscription auth is reusable by OpenCode, Cline, and other harnesses. Some third-party restrictions apply on Max tokens.
- **Claude Max 5x** — $100/mo · Claude Sonnet 4.6, Opus 4.6 · BYOH. ~225 Sonnet messages per 5-hour window (5× Claude Pro). Claude Code subscription auth reusable by OpenCode, Cline, and other harnesses — though Anthropic restricts some third-party consumers of Max tokens.
- **Claude Max 20x** — $200/mo · Claude Sonnet 4.6, Opus 4.6 · BYOH. ~900 Sonnet messages per 5-hour window (20× Claude Pro). Heaviest Claude Code tier for full-day Opus runs. Subscription auth reusable by OpenCode / Cline with some third-party restrictions.
- **ChatGPT Plus** — $20/mo · GPT-5, o3, o4-mini, GPT-5 Codex · BYOH. Baseline Codex tier. Token-based 5-hour windows (since Apr 2 2026). Codex OAuth works in OpenCode, ForgeCode, Cline and other OpenAI-compatible harnesses.
- **ChatGPT Pro ($100)** — $100/mo · GPT-5, o3, o4-mini, GPT-5 Codex · BYOH. New $100 Pro tier (Apr 9 2026). 5× Codex of Plus — built for heavy agentic coding sessions. Same model access as $200 Pro. 10× boost through May 31 promo.
- **ChatGPT Pro ($200)** — $200/mo · GPT-5, o3, o4-mini, GPT-5 Codex · BYOH. Maximum Codex throughput — 20× Plus limits. Unlimited GPT-5 Instant/Thinking, 250 Deep Research runs/mo. Same harness OAuth as other tiers.

## Tracked hosts

- [Anthropic](https://www.anthropic.com/pricing): live pricing source
- [OpenAI](https://openai.com/api/pricing): live pricing source
- [Google AI Studio](https://ai.google.dev/pricing): live pricing source
- [Fireworks AI](https://fireworks.ai/pricing): live pricing source
- [Groq](https://groq.com/pricing): live pricing source
- [DeepSeek](https://api-docs.deepseek.com/quick_start/pricing): live pricing source
- [Cerebras](https://cloud.cerebras.ai/pricing): live pricing source
- [Minimax](): live pricing source
- [Amazon Bedrock](https://aws.amazon.com/bedrock/pricing/): live pricing source
- [Inceptron](): live pricing source
- [TensorX](https://tensorx.ai/models/): live pricing source
- [AKI.IO](https://aki.io/): live pricing source
- [Scaleway](https://www.scaleway.com/en/generative-apis/): live pricing source
- [Regolo](https://regolo.ai/pricing/): live pricing source
- [Lyceum Technology](https://lyceum.technology/): live pricing source
- [IONOS AI Model Hub](https://cloud.ionos.com/managed/ai-model-hub): live pricing source

## Disclaimer

The Token Bill is an independent, non-commercial price directory with no affiliate relationships and no advertising. Provider names, model names, and trademarks are the property of their respective owners and are used nominatively only. Prices are best-effort — always verify on the provider's own page before buying. Nothing here is financial or purchasing advice.