Third-party resale listings
A snapshot of third-party resale prices for LLM APIs, captured 2026-09-22 from the reseller marketplace subrouter.ai's public pricing list. These are not official provider prices — see each model's own page for the full listing table and per-reseller caveats.
How this data is filtered. Every reseller row is checked
against this site's own official price for the same model (from Artificial Analysis) before it's
shown: it must be a plain per-token rate (not a subscription/tiered/per-call plan), it must come
from a "provider" or "official" listing group (pooled "shared-plan" rows are excluded), the
model must have a known official price to compare against, its output/input price ratio can't
differ from the official ratio by more than 3×, its blended price can't be
implausibly low or more than 5× the official blended price, and only the cheapest row
per reseller per model is kept. Rows priced far below official (under 10% of it) are excluded
from the "cheapest trusted price" figure but still shown, flagged, in each model's full table.
Coverage
Snapshot dated 2026-09-22.
- 81 models on this site have at least one resale listing
- 746 total reseller listings kept after filtering
- Median resale price is 9% of official price
Discarded rows by reason
Transparency: every reseller row this build looked at but did not show, and why.
| Reason | Count |
|---|---|
| below_floor | 10 |
| duplicate_provider | 18 |
| group_excluded | 1 |
| no_official_price | 21 |
| not_ratio_or_nonpositive | 131 |
| ratio_mismatch | 8 |
Most-listed models
Top 20 of 81 models by number of resellers listed.
| Model | Provider | Resellers |
|---|---|---|
| GPT-5.6 Sol (max) | OpenAI | 41 |
| Claude Opus 4.8 (Adaptive Reasoning, Max Effort) | Anthropic | 34 |
| Claude Fable 5 (Adaptive Reasoning, Max Effort, Opus 4.8 Fallback) | Anthropic | 33 |
| GPT-5.5 (xhigh) | OpenAI | 33 |
| GPT-5.6 Terra (max) | OpenAI | 33 |
| Claude Opus 5 (Adaptive Reasoning, Max Effort) | Anthropic | 31 |
| Claude Opus 4.7 (Adaptive Reasoning, Max Effort) | Anthropic | 30 |
| DeepSeek V4 Pro 0813 (Reasoning, Max Effort) | DeepSeek | 30 |
| Claude Opus 4.6 (Non-reasoning, High Effort) | Anthropic | 29 |
| DeepSeek V4 Flash 0731 (Reasoning, Max Effort) | DeepSeek | 28 |
| Claude Sonnet 4.6 (Non-reasoning, High Effort) | Anthropic | 27 |
| GPT-6 Astra (max) | OpenAI | 27 |
| Claude Sonnet 5 (Adaptive Reasoning, Max Effort) | Anthropic | 21 |
| Kimi K3 (max) | Kimi | 21 |
| DeepSeek V4.1 Flash (Reasoning, Max Effort) | DeepSeek | 18 |
| GLM-5.3 (max) | Z AI | 17 |
| GLM 5.3 Flash | Z AI | 16 |
| Grok 4.6 (high) | SpaceXAI | 16 |
| GPT-5.6 Luna (max) | OpenAI | 15 |
| Claude Fable 5.1 (Adaptive Reasoning, Max Effort, Default Fallback) | Anthropic | 14 |
Single-reseller models
12 models with exactly one reseller passing our checks.
| Model | Provider |
|---|---|
| DeepSeek V3 0324 | DeepSeek |
| GLM-4.7 (Reasoning) | Z AI |
| GLM-5 (Reasoning) | Z AI |
| GPT-3.5 Turbo | OpenAI |
| GPT-4.1 | OpenAI |
| GPT-4.1 mini | OpenAI |
| GPT-4.1 nano | OpenAI |
| GPT-4o (Nov '24) | OpenAI |
| GPT-4o mini | OpenAI |
| Gemini 3 Flash Preview (Non-reasoning) | |
| Kimi K2 | Kimi |
| MiniMax-M2 | MiniMax |
Resale listings: subrouter.ai public marketplace snapshot 2026-09-22. LLMAPIProviders has no affiliation with any listed reseller — evaluate them yourself before committing spend.