# Compare x402 AI and inference APIs x402 endpoints that run a model for you — LLM inference, embeddings, generation and agent tooling — compared on price per call, network and asset. > These are observations, not accusations. A band of LOW, MEDIUM or HIGH describes how much of > what we check we were able to confirm — it is not a judgement about the operator of an > endpoint. Every signal, its weight and its rationale are published at /methodology, and any > operator can claim an endpoint, correct a fact or opt out at /crawler. | Endpoint | Price per call | Network | Asset | Pays to | Observed signals | Watched for | Terms last observed | | -------------------------------- | ---------------- | ---------------- | ---------------- | ------------------------------------------ | ---------------------------- | --------------------- | -------------------- | | GEDX402 Chat | 0.001000 USDC | eip155:8453 | USDC | 0xb3C2776cE3F99cb3366520c27b4aC5D436942Ab6 | 96/100 LOW (methodology v1) | under a day, 1 probe | 2026-08-15T04:45:38Z | | 2s | 0.002500 USDC | eip155:8453 | USDC | 0x2b6D4988Db4723E6908Db86Ab2b8dFBc51FC32C5 | 100/100 LOW (methodology v1) | under a day, 1 probe | 2026-08-15T05:16:01Z | | 2s | 0.002500 USDC | eip155:8453 | USDC | 0x2b6D4988Db4723E6908Db86Ab2b8dFBc51FC32C5 | 100/100 LOW (methodology v1) | under a day, 1 probe | 2026-08-15T06:01:04Z | | 2s | 0.002500 USDC | eip155:8453 | USDC | 0x2b6D4988Db4723E6908Db86Ab2b8dFBc51FC32C5 | 100/100 LOW (methodology v1) | under a day, 1 probe | 2026-08-15T05:00:55Z | | 2s | 0.002500 USDC | eip155:8453 | USDC | 0x2b6D4988Db4723E6908Db86Ab2b8dFBc51FC32C5 | 100/100 LOW (methodology v1) | under a day, 2 probes | 2026-08-15T06:15:46Z | | 2s | 0.002500 USDC | eip155:8453 | USDC | 0x2b6D4988Db4723E6908Db86Ab2b8dFBc51FC32C5 | 100/100 LOW (methodology v1) | under a day, 1 probe | 2026-08-15T05:15:45Z | | Stock Trends Market Intelligence | 0.002500 USDC | eip155:8453 | USDC | 0xAEeb8EaBdC05532a26123045f99D208EB1cC00ab | 96/100 LOW (methodology v1) | under a day, 1 probe | 2026-08-15T05:00:57Z | | 2s | 0.003000 USDC | eip155:8453 | USDC | 0x2b6D4988Db4723E6908Db86Ab2b8dFBc51FC32C5 | 100/100 LOW (methodology v1) | under a day, 1 probe | 2026-08-15T06:00:48Z | | Stock Trends Market Intelligence | 0.003500 USDC | eip155:8453 | USDC | 0xAEeb8EaBdC05532a26123045f99D208EB1cC00ab | 96/100 LOW (methodology v1) | under a day, 1 probe | 2026-08-15T06:01:02Z | | 2s | 0.004500 USDC | eip155:8453 | USDC | 0x2b6D4988Db4723E6908Db86Ab2b8dFBc51FC32C5 | 100/100 LOW (methodology v1) | under a day, 1 probe | 2026-08-15T06:01:05Z | | 2s | 0.004500 USDC | eip155:8453 | USDC | 0x2b6D4988Db4723E6908Db86Ab2b8dFBc51FC32C5 | 100/100 LOW (methodology v1) | under a day, 1 probe | 2026-08-15T06:00:55Z | | GEDX402 Chat | 0.005000 USDC | eip155:8453 | USDC | 0xb3C2776cE3F99cb3366520c27b4aC5D436942Ab6 | 96/100 LOW (methodology v1) | under a day, 1 probe | 2026-08-15T05:46:01Z | | Quartermaster Web Extract | 0.005000 USDC | eip155:8453 | USDC | 0x5253666c40432f009ebd99908b9cd2a7fd1019cb | 96/100 LOW (methodology v1) | under a day, 1 probe | 2026-08-15T06:00:20Z | | DeepSeek | 0.007500 USDC | eip155:8453 | USDC | 0x3C74Ce3f8566Ff04B312320E7c264A691c409992 | 96/100 LOW (methodology v1) | under a day, 1 probe | 2026-08-15T05:01:01Z | | Trust Index | 0.010000 USDC | eip155:8453 | USDC | 0x53831fab63ba75a958115c3b4eb598a10cfcc7e0 | 96/100 LOW (methodology v1) | under a day, 1 probe | 2026-08-15T06:00:34Z | | fresh-feeds | 0.020000 USDC | eip155:8453 | USDC | 0x63347CD326a50696F76cfE4B8Ce079D3c8ED3039 | 96/100 LOW (methodology v1) | under a day, 1 probe | 2026-08-15T06:00:45Z | | CheapTokens AI Inference | 1.000000 USDC | eip155:8453 | USDC | 0x213c8d7434E2ae7AA1C392767c5120778D413215 | 96/100 LOW (methodology v1) | under a day, 1 probe | 2026-08-15T04:30:50Z | | Exa | — not observed | — not observed | — not observed | — not observed | — not observed | under a day, 1 probe | 2026-08-15T04:45:17Z | | svgsmith | — not observed | — not observed | — not observed | — not observed | — not observed | under a day, 2 probes | 2026-08-15T14:31:16Z | | X402 Data API | — not observed | — not observed | — not observed | — not observed | — not observed | under a day, 2 probes | 2026-08-15T14:31:22Z | | X402 Data API | — not observed | — not observed | — not observed | — not observed | — not observed | under a day, 2 probes | 2026-08-15T14:31:23Z | | X402 Data API | — not observed | — not observed | — not observed | — not observed | — not observed | under a day, 2 probes | 2026-08-15T14:31:21Z | | BlockRun.AI | — not probed yet | — not probed yet | — not probed yet | — not probed yet | — not probed yet | — not probed yet | — not probed yet | | BlockRun.AI | — not probed yet | — not probed yet | — not probed yet | — not probed yet | — not probed yet | — not probed yet | — not probed yet | | MiroShark | — not probed yet | — not probed yet | — not probed yet | — not probed yet | — not probed yet | — not probed yet | — not probed yet | Sorted by price. ## What the facilitator claims These figures are published by the facilitator that lists the endpoint. They are not our observations, and they are shown separately from them for that reason. An advertised price and the price the endpoint actually served us are two different facts, and a disagreement between them is a finding rather than an error. | Endpoint | Listed by | Usage (30d) | Unique payers (30d) | Advertised price | | -------------------------------- | ------------ | -------------------- | ------------------- | ---------------- | | GEDX402 Chat | coinbase-cdp | 219 calls in 30 days | 10 unique payers | 1000 atomic | | 2s | coinbase-cdp | 146 calls in 30 days | 85 unique payers | 2500 atomic | | 2s | coinbase-cdp | 70 calls in 30 days | 56 unique payers | 2500 atomic | | 2s | coinbase-cdp | 405 calls in 30 days | 109 unique payers | 2500 atomic | | 2s | coinbase-cdp | 87 calls in 30 days | 66 unique payers | 2500 atomic | | 2s | coinbase-cdp | 170 calls in 30 days | 96 unique payers | 2500 atomic | | Stock Trends Market Intelligence | coinbase-cdp | 22 calls in 30 days | 7 unique payers | 2500 atomic | | 2s | coinbase-cdp | 81 calls in 30 days | 60 unique payers | 3000 atomic | | Stock Trends Market Intelligence | coinbase-cdp | 14 calls in 30 days | 3 unique payers | 3500 atomic | | 2s | coinbase-cdp | 69 calls in 30 days | 56 unique payers | 4500 atomic | | 2s | coinbase-cdp | 80 calls in 30 days | 57 unique payers | 4500 atomic | | GEDX402 Chat | coinbase-cdp | 5 calls in 30 days | 4 unique payers | 5000 atomic | | Quartermaster Web Extract | coinbase-cdp | 23 calls in 30 days | 3 unique payers | 5000 atomic | | DeepSeek | coinbase-cdp | 55 calls in 30 days | 6 unique payers | 2500 atomic | | Trust Index | coinbase-cdp | 4 calls in 30 days | 4 unique payers | 10000 atomic | | fresh-feeds | coinbase-cdp | 4 calls in 30 days | 4 unique payers | 20000 atomic | | CheapTokens AI Inference | coinbase-cdp | 164 calls in 30 days | 19 unique payers | 1000000 atomic | | Exa | coinbase-cdp | 378 calls in 30 days | 25 unique payers | 1000 atomic | | svgsmith | coinbase-cdp | 26 calls in 30 days | 3 unique payers | 20000 atomic | | X402 Data API | coinbase-cdp | 15 calls in 30 days | 9 unique payers | 1000000 atomic | | X402 Data API | coinbase-cdp | 11 calls in 30 days | 3 unique payers | 2000 atomic | | X402 Data API | coinbase-cdp | 9 calls in 30 days | 3 unique payers | 1000 atomic | | BlockRun.AI | coinbase-cdp | 675 calls in 30 days | 31 unique payers | 3000 atomic | | BlockRun.AI | coinbase-cdp | 51 calls in 30 days | 28 unique payers | 8500 atomic | | MiroShark | coinbase-cdp | 28 calls in 30 days | 11 unique payers | 1000000 atomic | ## What this table does and does not say - 3 of 25 endpoints here have no observed terms yet. Each says which state it is in — not in our index, not probed yet, no x402 challenge served, or did not answer — and none of them is shown as a zero. - Availability and latency are not shown here. They are aggregates over the sampled probe dataset and are read one endpoint at a time, which is what the history page does — this page reads the index and makes no outbound request. Estimating them from the probe records kept in the database instead would over-report failure, because routine successful probes are deliberately not kept. - Usage figures on this page are published by the facilitator that lists the endpoint (coinbase-cdp), not measured by us. They are shown because they are better data than anything we can collect — and attributed because they are somebody else's. - The facilitator's listing and the challenge we observed disagree about the price of deepseek.x402.press: the listing advertises 2500 and the endpoint served 7500 atomic units. Both are recorded; the amount we observed is what the endpoint actually asked us for. --- JSON: `/api/v1/compare?category=ai-inference` · full per-endpoint records: `/api/v1/endpoints` · categories: `/api/v1/categories` Every figure here is something we observed by asking the endpoint, except where it is labelled as a facilitator's claim. Nothing on this page is a statement about an operator.