OpenRouter API listing

Meta: Llama 3.3 70B Instruct (routed listing) API pricing on OpenRouter

OpenRouter's published API listing for Meta: Llama 3.3 70B Instruct (routed listing), with native prices and dated evidence. Check the official source for current availability and account eligibility.

Headline USD token rates per 1M; other currencies and units remain native. Original rates are below. Unknown means no unambiguous headline; inspect all rates and conditions below.
API listingInputCache readOutputCapability (Epoch AI)
Meta: Llama 3.3 70B Instruct (routed listing)USD 0.1 / 1M tokensrouted catalogUnknown — see conditionsUSD 0.32 / 1M tokensrouted catalogNo independent capability score

Rates, billing conditions and sources

Meta: Llama 3.3 70B Instruct (routed listing)

Service ID: meta-llama/llama-3.3-70b-instruct · Context window: 131K tokens · Listing status: active

All rates, conditions and sources

Input: USD 0.0000001 / 1 token · routed catalog · Modality: text

Region: Unknown · Endpoint scope: unknown · Tier: routed catalog · Billing: Unknown · Prompt bracket: none stated · Cache TTL: Unknown seconds · Valid through: Unknown

  • OpenRouter lowest listed endpoint price: DeepInfra (fp8, endpoint deepinfra/turbo). It is lowest on both input and output among the default-tier endpoints OpenRouter's official endpoints API lists for this model (opt-in flex, priority and ultrafast tier endpoints are left out), and a request with provider.sort = "price" tries endpoints in price order. OpenRouter's default routing load-balances across the lowest-cost stable providers weighted by inverse square of price, so a default request may be served, and billed, by a higher-priced endpoint.
  • Native USD per token as OpenRouter lists it, billed at the serving endpoint's rates. No discount or other charge is listed for this endpoint; free, flex, batch and priority variants are outside this reviewed scope.
  • Endpoint availability, status and regional or account eligibility are unverified (OpenRouter does not document its endpoint status codes). Listing is not an inference service test.
  • Listed uncached input or output token rate of that endpoint; unlisted cache prices are unknown, not zero.
  • Official model listing: https://openrouter.ai/meta-llama/llama-3.3-70b-instruct

Output: USD 0.00000032 / 1 token · routed catalog · Modality: text

Region: Unknown · Endpoint scope: unknown · Tier: routed catalog · Billing: Unknown · Prompt bracket: none stated · Cache TTL: Unknown seconds · Valid through: Unknown

  • OpenRouter lowest listed endpoint price: DeepInfra (fp8, endpoint deepinfra/turbo). It is lowest on both input and output among the default-tier endpoints OpenRouter's official endpoints API lists for this model (opt-in flex, priority and ultrafast tier endpoints are left out), and a request with provider.sort = "price" tries endpoints in price order. OpenRouter's default routing load-balances across the lowest-cost stable providers weighted by inverse square of price, so a default request may be served, and billed, by a higher-priced endpoint.
  • Native USD per token as OpenRouter lists it, billed at the serving endpoint's rates. No discount or other charge is listed for this endpoint; free, flex, batch and priority variants are outside this reviewed scope.
  • Endpoint availability, status and regional or account eligibility are unverified (OpenRouter does not document its endpoint status codes). Listing is not an inference service test.
  • Listed uncached input or output token rate of that endpoint; unlisted cache prices are unknown, not zero.
  • Official model listing: https://openrouter.ai/meta-llama/llama-3.3-70b-instruct

Developer: Unknown / not verified · API seller: OpenRouter

Checkpoint: Unknown / not verified · Lineage: Unknown / not verified

Official pricing ↗

Price observed: 2026-11-03T04:29:37.203Z · Price source checked: 2026-11-03T04:29:37.203Z · Price freshness: fresh

Weight-license reviewed: Unknown. Open weights do not establish permission for every commercial use; consult the linked terms.

Gating: unknown · Code license: Unknown / not verified · API availability: unknown (dated observation, not a live service check)

Source, conditions and dates

Raw API evidence ↗ · vendor-api · Claims: openrouter-llama-3-3-70b-instruct: pricing · openrouter-deepseek-v3-1: pricing · openrouter-minimax-m2: pricing · Retrieved · Effective: Unknown · Checked: 2026-10-02T21:21:16.675Z · Reviewed: Unknown · Review after: Unknown · Freshness: unknown · Confidence: B

  • Raw SHA-256: bbcb7bdbbded22ebaabd9009c1df39f0e723d5e2d12667df3bbb46975e5cc5bd; collector openrouter-text-catalog-2; inventory independently replayed without changing observation dates.
  • Pricing semantics reviewed 2026-10-02: https://openrouter.ai/docs/guides/overview/models (USD per token, top-provider pricing and conditional overrides). Only three exact text-only listings with no overrides or additional fees are supported.
  • Anonymous full default-text catalog, total and pagination verified. HF IDs/descriptions do not establish checkpoint, developer, license or weight access; no such joins are published.

Raw API evidence ↗ · vendor-api · Claims: openrouter-llama-3-3-70b-instruct: pricing · Retrieved · Effective: Unknown · Checked: 2026-10-15T01:49:27.881Z · Reviewed: Unknown · Review after: 2026-10-25T01:49:27.881Z · Freshness: stale · Confidence: B

  • Raw SHA-256: 69035a91ad4cbd69dfc6eacfc89be55fdd0551a54cb3802a7d011bdb2d5f2933; collector openrouter-endpoints-1; daily collection independently replayed with zero source requests.
  • Lowest listed endpoint of 11 default-tier endpoints for meta-llama/llama-3.3-70b-instruct: DeepInfra (fp8, endpoint deepinfra/turbo), lowest on both input and output as listed.
  • Semantics reviewed 2026-10-15 (#515): OpenRouter's provider-routing guide (https://openrouter.ai/docs/guides/routing/provider-selection.md, Markdown SHA-256 1206e349a01dd137cfc18683eacf7a82bab44c75e5d65d88f0a86b2f212c6e47) states that its default routing load-balances across the lowest-cost stable providers weighted by inverse square of price, and that provider.sort = "price" tries providers in price order; its service-tiers guide (https://openrouter.ai/docs/guides/features/service-tiers.md, Markdown SHA-256 76af5150365778a4775a4218de182b40665fbfd8c8b6babbd623e493b3d84a51) states that flex, priority and ultrafast tier endpoints (tier-suffixed tags) stay opt-in and that provider.sort does not opt into any tier. Its OpenAPI spec (https://openrouter.ai/docs/openapi/openapi.yaml, SHA-256 f36da4e53cd2f259791d0a2110bd1c3b4dd6fb3a1701bc9c6375e79c7df88ed0) defines per-endpoint USD-per-token prices and a discount fraction, and its provider integration guide (https://openrouter.ai/docs/guides/providers/integration.md, Markdown SHA-256 535c7594d2d5197584edf03fb8e4abb4afc64d7beccef929fdf9ed15f58c2224) states that a provider discount is applied to the displayed price users see and pay (user price = base price × (1 - discount_to_user)).
  • Endpoint status codes are undocumented and not used; no availability is asserted. HF IDs and descriptions do not establish checkpoint, developer, license or weight access; no such joins are published.

Raw API evidence ↗ · vendor-api · Claims: openrouter-llama-3-3-70b-instruct: pricing · Retrieved · Effective: Unknown · Checked: 2026-10-16T04:36:53.662Z · Reviewed: Unknown · Review after: 2026-10-26T04:36:53.662Z · Freshness: stale · Confidence: B

  • Raw SHA-256: 732a74c564eb4c9f526e2d71362d4f5f4233c0588a421fb2d8abf73589d76942; collector openrouter-endpoints-1; daily collection independently replayed with zero source requests.
  • Lowest listed endpoint of 11 default-tier endpoints for meta-llama/llama-3.3-70b-instruct: DeepInfra (fp8, endpoint deepinfra/turbo), lowest on both input and output as listed.
  • Semantics reviewed 2026-10-15 (#515): OpenRouter's provider-routing guide (https://openrouter.ai/docs/guides/routing/provider-selection.md, Markdown SHA-256 1206e349a01dd137cfc18683eacf7a82bab44c75e5d65d88f0a86b2f212c6e47) states that its default routing load-balances across the lowest-cost stable providers weighted by inverse square of price, and that provider.sort = "price" tries providers in price order; its service-tiers guide (https://openrouter.ai/docs/guides/features/service-tiers.md, Markdown SHA-256 76af5150365778a4775a4218de182b40665fbfd8c8b6babbd623e493b3d84a51) states that flex, priority and ultrafast tier endpoints (tier-suffixed tags) stay opt-in and that provider.sort does not opt into any tier. Its OpenAPI spec (https://openrouter.ai/docs/openapi/openapi.yaml, SHA-256 f36da4e53cd2f259791d0a2110bd1c3b4dd6fb3a1701bc9c6375e79c7df88ed0) defines per-endpoint USD-per-token prices and a discount fraction, and its provider integration guide (https://openrouter.ai/docs/guides/providers/integration.md, Markdown SHA-256 535c7594d2d5197584edf03fb8e4abb4afc64d7beccef929fdf9ed15f58c2224) states that a provider discount is applied to the displayed price users see and pay (user price = base price × (1 - discount_to_user)).
  • Endpoint status codes are undocumented and not used; no availability is asserted. HF IDs and descriptions do not establish checkpoint, developer, license or weight access; no such joins are published.

Raw API evidence ↗ · vendor-api · Claims: openrouter-llama-3-3-70b-instruct: pricing · Retrieved · Effective: Unknown · Checked: 2026-10-17T04:36:06.688Z · Reviewed: Unknown · Review after: 2026-10-27T04:36:06.688Z · Freshness: stale · Confidence: B

  • Raw SHA-256: 2e3e7a557c8d4ffa1ce7add34efa7a7155ba7d9028f87357b43a922a161b7698; collector openrouter-endpoints-1; daily collection independently replayed with zero source requests.
  • Lowest listed endpoint of 11 default-tier endpoints for meta-llama/llama-3.3-70b-instruct: DeepInfra (fp8, endpoint deepinfra/turbo), lowest on both input and output as listed.
  • Semantics reviewed 2026-10-15 (#515): OpenRouter's provider-routing guide (https://openrouter.ai/docs/guides/routing/provider-selection.md, Markdown SHA-256 1206e349a01dd137cfc18683eacf7a82bab44c75e5d65d88f0a86b2f212c6e47) states that its default routing load-balances across the lowest-cost stable providers weighted by inverse square of price, and that provider.sort = "price" tries providers in price order; its service-tiers guide (https://openrouter.ai/docs/guides/features/service-tiers.md, Markdown SHA-256 76af5150365778a4775a4218de182b40665fbfd8c8b6babbd623e493b3d84a51) states that flex, priority and ultrafast tier endpoints (tier-suffixed tags) stay opt-in and that provider.sort does not opt into any tier. Its OpenAPI spec (https://openrouter.ai/docs/openapi/openapi.yaml, SHA-256 f36da4e53cd2f259791d0a2110bd1c3b4dd6fb3a1701bc9c6375e79c7df88ed0) defines per-endpoint USD-per-token prices and a discount fraction, and its provider integration guide (https://openrouter.ai/docs/guides/providers/integration.md, Markdown SHA-256 535c7594d2d5197584edf03fb8e4abb4afc64d7beccef929fdf9ed15f58c2224) states that a provider discount is applied to the displayed price users see and pay (user price = base price × (1 - discount_to_user)).
  • Endpoint status codes are undocumented and not used; no availability is asserted. HF IDs and descriptions do not establish checkpoint, developer, license or weight access; no such joins are published.

Raw API evidence ↗ · vendor-api · Claims: openrouter-llama-3-3-70b-instruct: pricing · Retrieved · Effective: Unknown · Checked: 2026-10-18T04:36:34.202Z · Reviewed: Unknown · Review after: 2026-10-28T04:36:34.202Z · Freshness: stale · Confidence: B

  • Raw SHA-256: a103d48973eaa975b2bdd56a3e0e2bf73a4b71075d4404f099ada95e187b09f5; collector openrouter-endpoints-1; daily collection independently replayed with zero source requests.
  • Lowest listed endpoint of 11 default-tier endpoints for meta-llama/llama-3.3-70b-instruct: DeepInfra (fp8, endpoint deepinfra/turbo), lowest on both input and output as listed.
  • Semantics reviewed 2026-10-15 (#515): OpenRouter's provider-routing guide (https://openrouter.ai/docs/guides/routing/provider-selection.md, Markdown SHA-256 1206e349a01dd137cfc18683eacf7a82bab44c75e5d65d88f0a86b2f212c6e47) states that its default routing load-balances across the lowest-cost stable providers weighted by inverse square of price, and that provider.sort = "price" tries providers in price order; its service-tiers guide (https://openrouter.ai/docs/guides/features/service-tiers.md, Markdown SHA-256 76af5150365778a4775a4218de182b40665fbfd8c8b6babbd623e493b3d84a51) states that flex, priority and ultrafast tier endpoints (tier-suffixed tags) stay opt-in and that provider.sort does not opt into any tier. Its OpenAPI spec (https://openrouter.ai/docs/openapi/openapi.yaml, SHA-256 f36da4e53cd2f259791d0a2110bd1c3b4dd6fb3a1701bc9c6375e79c7df88ed0) defines per-endpoint USD-per-token prices and a discount fraction, and its provider integration guide (https://openrouter.ai/docs/guides/providers/integration.md, Markdown SHA-256 535c7594d2d5197584edf03fb8e4abb4afc64d7beccef929fdf9ed15f58c2224) states that a provider discount is applied to the displayed price users see and pay (user price = base price × (1 - discount_to_user)).
  • Endpoint status codes are undocumented and not used; no availability is asserted. HF IDs and descriptions do not establish checkpoint, developer, license or weight access; no such joins are published.

Raw API evidence ↗ · vendor-api · Claims: openrouter-llama-3-3-70b-instruct: pricing · Retrieved · Effective: Unknown · Checked: 2026-10-19T04:41:45.805Z · Reviewed: Unknown · Review after: 2026-10-29T04:41:45.805Z · Freshness: stale · Confidence: B

  • Raw SHA-256: 50c375a9563e6be559892680edb5c93b3305dcf67a1b058ffbe630539381ade8; collector openrouter-endpoints-1; daily collection independently replayed with zero source requests.
  • Lowest listed endpoint of 11 default-tier endpoints for meta-llama/llama-3.3-70b-instruct: DeepInfra (fp8, endpoint deepinfra/turbo), lowest on both input and output as listed.
  • Semantics reviewed 2026-10-15 (#515): OpenRouter's provider-routing guide (https://openrouter.ai/docs/guides/routing/provider-selection.md, Markdown SHA-256 1206e349a01dd137cfc18683eacf7a82bab44c75e5d65d88f0a86b2f212c6e47) states that its default routing load-balances across the lowest-cost stable providers weighted by inverse square of price, and that provider.sort = "price" tries providers in price order; its service-tiers guide (https://openrouter.ai/docs/guides/features/service-tiers.md, Markdown SHA-256 76af5150365778a4775a4218de182b40665fbfd8c8b6babbd623e493b3d84a51) states that flex, priority and ultrafast tier endpoints (tier-suffixed tags) stay opt-in and that provider.sort does not opt into any tier. Its OpenAPI spec (https://openrouter.ai/docs/openapi/openapi.yaml, SHA-256 f36da4e53cd2f259791d0a2110bd1c3b4dd6fb3a1701bc9c6375e79c7df88ed0) defines per-endpoint USD-per-token prices and a discount fraction, and its provider integration guide (https://openrouter.ai/docs/guides/providers/integration.md, Markdown SHA-256 535c7594d2d5197584edf03fb8e4abb4afc64d7beccef929fdf9ed15f58c2224) states that a provider discount is applied to the displayed price users see and pay (user price = base price × (1 - discount_to_user)).
  • Endpoint status codes are undocumented and not used; no availability is asserted. HF IDs and descriptions do not establish checkpoint, developer, license or weight access; no such joins are published.

Raw API evidence ↗ · vendor-api · Claims: openrouter-llama-3-3-70b-instruct: pricing · Retrieved · Effective: Unknown · Checked: 2026-10-20T04:38:13.051Z · Reviewed: Unknown · Review after: 2026-10-30T04:38:13.051Z · Freshness: stale · Confidence: B

  • Raw SHA-256: aa7ef0e7fb4ac8626d8f586263f70ae34338b8a7eb7008ee6e4210d3e24ed6b3; collector openrouter-endpoints-1; daily collection independently replayed with zero source requests.
  • Lowest listed endpoint of 11 default-tier endpoints for meta-llama/llama-3.3-70b-instruct: DeepInfra (fp8, endpoint deepinfra/turbo), lowest on both input and output as listed.
  • Semantics reviewed 2026-10-15 (#515): OpenRouter's provider-routing guide (https://openrouter.ai/docs/guides/routing/provider-selection.md, Markdown SHA-256 1206e349a01dd137cfc18683eacf7a82bab44c75e5d65d88f0a86b2f212c6e47) states that its default routing load-balances across the lowest-cost stable providers weighted by inverse square of price, and that provider.sort = "price" tries providers in price order; its service-tiers guide (https://openrouter.ai/docs/guides/features/service-tiers.md, Markdown SHA-256 76af5150365778a4775a4218de182b40665fbfd8c8b6babbd623e493b3d84a51) states that flex, priority and ultrafast tier endpoints (tier-suffixed tags) stay opt-in and that provider.sort does not opt into any tier. Its OpenAPI spec (https://openrouter.ai/docs/openapi/openapi.yaml, SHA-256 f36da4e53cd2f259791d0a2110bd1c3b4dd6fb3a1701bc9c6375e79c7df88ed0) defines per-endpoint USD-per-token prices and a discount fraction, and its provider integration guide (https://openrouter.ai/docs/guides/providers/integration.md, Markdown SHA-256 535c7594d2d5197584edf03fb8e4abb4afc64d7beccef929fdf9ed15f58c2224) states that a provider discount is applied to the displayed price users see and pay (user price = base price × (1 - discount_to_user)).
  • Endpoint status codes are undocumented and not used; no availability is asserted. HF IDs and descriptions do not establish checkpoint, developer, license or weight access; no such joins are published.

Raw API evidence ↗ · vendor-api · Claims: openrouter-llama-3-3-70b-instruct: pricing · Retrieved · Effective: Unknown · Checked: 2026-10-25T04:40:32.150Z · Reviewed: Unknown · Review after: 2026-11-04T04:40:32.150Z · Freshness: stale · Confidence: B

  • Raw SHA-256: 85f169cacb0eeca83a140cd871526f5f0a472296cb4236e4342840329b30f030; collector openrouter-endpoints-1; daily collection independently replayed with zero source requests.
  • Lowest listed endpoint of 10 default-tier endpoints for meta-llama/llama-3.3-70b-instruct: DeepInfra (fp8, endpoint deepinfra/turbo), lowest on both input and output as listed.
  • Semantics reviewed 2026-10-15 (#515): OpenRouter's provider-routing guide (https://openrouter.ai/docs/guides/routing/provider-selection.md, Markdown SHA-256 1206e349a01dd137cfc18683eacf7a82bab44c75e5d65d88f0a86b2f212c6e47) states that its default routing load-balances across the lowest-cost stable providers weighted by inverse square of price, and that provider.sort = "price" tries providers in price order; its service-tiers guide (https://openrouter.ai/docs/guides/features/service-tiers.md, Markdown SHA-256 76af5150365778a4775a4218de182b40665fbfd8c8b6babbd623e493b3d84a51) states that flex, priority and ultrafast tier endpoints (tier-suffixed tags) stay opt-in and that provider.sort does not opt into any tier. Its OpenAPI spec (https://openrouter.ai/docs/openapi/openapi.yaml, SHA-256 f36da4e53cd2f259791d0a2110bd1c3b4dd6fb3a1701bc9c6375e79c7df88ed0) defines per-endpoint USD-per-token prices and a discount fraction, and its provider integration guide (https://openrouter.ai/docs/guides/providers/integration.md, Markdown SHA-256 535c7594d2d5197584edf03fb8e4abb4afc64d7beccef929fdf9ed15f58c2224) states that a provider discount is applied to the displayed price users see and pay (user price = base price × (1 - discount_to_user)).
  • Endpoint status codes are undocumented and not used; no availability is asserted. HF IDs and descriptions do not establish checkpoint, developer, license or weight access; no such joins are published.

Raw API evidence ↗ · vendor-api · Claims: openrouter-llama-3-3-70b-instruct: pricing · Retrieved · Effective: Unknown · Checked: 2026-10-26T01:03:41.180Z · Reviewed: Unknown · Review after: 2026-11-05T01:03:41.180Z · Freshness: stale · Confidence: B

  • Raw SHA-256: 466a9c2c313fb54d938715acf13cf1bb42c58d3db86596daaf02e0e928758517; collector openrouter-endpoints-1; daily collection independently replayed with zero source requests.
  • Lowest listed endpoint of 10 default-tier endpoints for meta-llama/llama-3.3-70b-instruct: DeepInfra (fp8, endpoint deepinfra/turbo), lowest on both input and output as listed.
  • Semantics reviewed 2026-10-15 (#515): OpenRouter's provider-routing guide (https://openrouter.ai/docs/guides/routing/provider-selection.md, Markdown SHA-256 1206e349a01dd137cfc18683eacf7a82bab44c75e5d65d88f0a86b2f212c6e47) states that its default routing load-balances across the lowest-cost stable providers weighted by inverse square of price, and that provider.sort = "price" tries providers in price order; its service-tiers guide (https://openrouter.ai/docs/guides/features/service-tiers.md, Markdown SHA-256 76af5150365778a4775a4218de182b40665fbfd8c8b6babbd623e493b3d84a51) states that flex, priority and ultrafast tier endpoints (tier-suffixed tags) stay opt-in and that provider.sort does not opt into any tier. Its OpenAPI spec (https://openrouter.ai/docs/openapi/openapi.yaml, SHA-256 f36da4e53cd2f259791d0a2110bd1c3b4dd6fb3a1701bc9c6375e79c7df88ed0) defines per-endpoint USD-per-token prices and a discount fraction, and its provider integration guide (https://openrouter.ai/docs/guides/providers/integration.md, Markdown SHA-256 535c7594d2d5197584edf03fb8e4abb4afc64d7beccef929fdf9ed15f58c2224) states that a provider discount is applied to the displayed price users see and pay (user price = base price × (1 - discount_to_user)).
  • Endpoint status codes are undocumented and not used; no availability is asserted. HF IDs and descriptions do not establish checkpoint, developer, license or weight access; no such joins are published.

Raw API evidence ↗ · vendor-api · Claims: openrouter-llama-3-3-70b-instruct: pricing · openrouter-llama-3-3-70b-instruct: specifications · Retrieved · Effective: Unknown · Checked: 2026-10-27T01:28:20.212Z · Reviewed: Unknown · Review after: 2026-11-06T01:28:20.212Z · Freshness: stale · Confidence: B

  • Raw SHA-256: aaab5667e41014c935ac0f73582c46a00ad14460106dace1f00f438087e5c2ee; collector openrouter-endpoints-1; daily collection independently replayed with zero source requests.
  • Lowest listed endpoint of 10 default-tier endpoints for meta-llama/llama-3.3-70b-instruct: DeepInfra (fp8, endpoint deepinfra/turbo), lowest on both input and output as listed.
  • Semantics reviewed 2026-10-15 (#515): OpenRouter's provider-routing guide (https://openrouter.ai/docs/guides/routing/provider-selection.md, Markdown SHA-256 1206e349a01dd137cfc18683eacf7a82bab44c75e5d65d88f0a86b2f212c6e47) states that its default routing load-balances across the lowest-cost stable providers weighted by inverse square of price, and that provider.sort = "price" tries providers in price order; its service-tiers guide (https://openrouter.ai/docs/guides/features/service-tiers.md, Markdown SHA-256 76af5150365778a4775a4218de182b40665fbfd8c8b6babbd623e493b3d84a51) states that flex, priority and ultrafast tier endpoints (tier-suffixed tags) stay opt-in and that provider.sort does not opt into any tier. Its OpenAPI spec (https://openrouter.ai/docs/openapi/openapi.yaml, SHA-256 f36da4e53cd2f259791d0a2110bd1c3b4dd6fb3a1701bc9c6375e79c7df88ed0) defines per-endpoint USD-per-token prices and a discount fraction, and its provider integration guide (https://openrouter.ai/docs/guides/providers/integration.md, Markdown SHA-256 535c7594d2d5197584edf03fb8e4abb4afc64d7beccef929fdf9ed15f58c2224) states that a provider discount is applied to the displayed price users see and pay (user price = base price × (1 - discount_to_user)).
  • Endpoint status codes are undocumented and not used; no availability is asserted. HF IDs and descriptions do not establish checkpoint, developer, license or weight access; no such joins are published.
  • Context window: 131072 tokens, the context_length OpenRouter lists for that endpoint. OpenRouter's provider integration guide (https://openrouter.ai/docs/guides/providers/integration.md, Markdown SHA-256 5881020f822b599ece83ff232645dc31c6ac2d1bd7e354473ebc425951b95742, fetched 2026-10-26T19:37:54Z) defines a provider's declared max_context_length (text) as "the total context window: input and output tokens combined", which "Becomes the endpoint's context length" ("Held if set manually": OpenRouter staff may set it by hand). The value follows the lowest-price endpoint; the model's developer and other endpoints may state different limits.

Raw API evidence ↗ · vendor-api · Claims: openrouter-llama-3-3-70b-instruct: pricing · openrouter-llama-3-3-70b-instruct: specifications · Retrieved · Effective: Unknown · Checked: 2026-10-28T04:31:55.028Z · Reviewed: Unknown · Review after: 2026-11-07T04:31:55.028Z · Freshness: stale · Confidence: B

  • Raw SHA-256: 6f723090b5b0c23f3a96eed564e1f2838037cea2070586dc867c9a3fa187c822; collector openrouter-endpoints-1; daily collection independently replayed with zero source requests.
  • Lowest listed endpoint of 10 default-tier endpoints for meta-llama/llama-3.3-70b-instruct: DeepInfra (fp8, endpoint deepinfra/turbo), lowest on both input and output as listed.
  • Semantics reviewed 2026-10-15 (#515): OpenRouter's provider-routing guide (https://openrouter.ai/docs/guides/routing/provider-selection.md, Markdown SHA-256 1206e349a01dd137cfc18683eacf7a82bab44c75e5d65d88f0a86b2f212c6e47) states that its default routing load-balances across the lowest-cost stable providers weighted by inverse square of price, and that provider.sort = "price" tries providers in price order; its service-tiers guide (https://openrouter.ai/docs/guides/features/service-tiers.md, Markdown SHA-256 76af5150365778a4775a4218de182b40665fbfd8c8b6babbd623e493b3d84a51) states that flex, priority and ultrafast tier endpoints (tier-suffixed tags) stay opt-in and that provider.sort does not opt into any tier. Its OpenAPI spec (https://openrouter.ai/docs/openapi/openapi.yaml, SHA-256 f36da4e53cd2f259791d0a2110bd1c3b4dd6fb3a1701bc9c6375e79c7df88ed0) defines per-endpoint USD-per-token prices and a discount fraction, and its provider integration guide (https://openrouter.ai/docs/guides/providers/integration.md, Markdown SHA-256 535c7594d2d5197584edf03fb8e4abb4afc64d7beccef929fdf9ed15f58c2224) states that a provider discount is applied to the displayed price users see and pay (user price = base price × (1 - discount_to_user)).
  • Endpoint status codes are undocumented and not used; no availability is asserted. HF IDs and descriptions do not establish checkpoint, developer, license or weight access; no such joins are published.
  • Context window: 131072 tokens, the context_length OpenRouter lists for that endpoint. OpenRouter's provider integration guide (https://openrouter.ai/docs/guides/providers/integration.md, Markdown SHA-256 5881020f822b599ece83ff232645dc31c6ac2d1bd7e354473ebc425951b95742, fetched 2026-10-26T19:37:54Z) defines a provider's declared max_context_length (text) as "the total context window: input and output tokens combined", which "Becomes the endpoint's context length" ("Held if set manually": OpenRouter staff may set it by hand). The value follows the lowest-price endpoint; the model's developer and other endpoints may state different limits.

Raw API evidence ↗ · vendor-api · Claims: openrouter-llama-3-3-70b-instruct: pricing · openrouter-llama-3-3-70b-instruct: specifications · Retrieved · Effective: Unknown · Checked: 2026-10-30T04:32:32.204Z · Reviewed: Unknown · Review after: 2026-11-09T04:32:32.204Z · Freshness: stale · Confidence: B

  • Raw SHA-256: b877d89635ea7fc9bf878b6b86ea449b7f620c2f6338df9f197b6dadc5b27dcf; collector openrouter-endpoints-1; daily collection independently replayed with zero source requests.
  • Lowest listed endpoint of 10 default-tier endpoints for meta-llama/llama-3.3-70b-instruct: DeepInfra (fp8, endpoint deepinfra/turbo), lowest on both input and output as listed.
  • Semantics reviewed 2026-10-15 (#515): OpenRouter's provider-routing guide (https://openrouter.ai/docs/guides/routing/provider-selection.md, Markdown SHA-256 1206e349a01dd137cfc18683eacf7a82bab44c75e5d65d88f0a86b2f212c6e47) states that its default routing load-balances across the lowest-cost stable providers weighted by inverse square of price, and that provider.sort = "price" tries providers in price order; its service-tiers guide (https://openrouter.ai/docs/guides/features/service-tiers.md, Markdown SHA-256 76af5150365778a4775a4218de182b40665fbfd8c8b6babbd623e493b3d84a51) states that flex, priority and ultrafast tier endpoints (tier-suffixed tags) stay opt-in and that provider.sort does not opt into any tier. Its OpenAPI spec (https://openrouter.ai/docs/openapi/openapi.yaml, SHA-256 f36da4e53cd2f259791d0a2110bd1c3b4dd6fb3a1701bc9c6375e79c7df88ed0) defines per-endpoint USD-per-token prices and a discount fraction, and its provider integration guide (https://openrouter.ai/docs/guides/providers/integration.md, Markdown SHA-256 535c7594d2d5197584edf03fb8e4abb4afc64d7beccef929fdf9ed15f58c2224) states that a provider discount is applied to the displayed price users see and pay (user price = base price × (1 - discount_to_user)).
  • Endpoint status codes are undocumented and not used; no availability is asserted. HF IDs and descriptions do not establish checkpoint, developer, license or weight access; no such joins are published.
  • Context window: 131072 tokens, the context_length OpenRouter lists for that endpoint. OpenRouter's provider integration guide (https://openrouter.ai/docs/guides/providers/integration.md, Markdown SHA-256 5881020f822b599ece83ff232645dc31c6ac2d1bd7e354473ebc425951b95742, fetched 2026-10-26T19:37:54Z) defines a provider's declared max_context_length (text) as "the total context window: input and output tokens combined", which "Becomes the endpoint's context length" ("Held if set manually": OpenRouter staff may set it by hand). The value follows the lowest-price endpoint; the model's developer and other endpoints may state different limits.

Raw API evidence ↗ · vendor-api · Claims: openrouter-llama-3-3-70b-instruct: pricing · openrouter-llama-3-3-70b-instruct: specifications · Retrieved · Effective: Unknown · Checked: 2026-11-01T04:29:05.689Z · Reviewed: Unknown · Review after: 2026-11-11T04:29:05.689Z · Freshness: stale · Confidence: B

  • Raw SHA-256: cacde41ba41b234a387581d39a5f9437b0eaf86ce86e01f5f1286622f6581115; collector openrouter-endpoints-1; daily collection independently replayed with zero source requests.
  • Lowest listed endpoint of 10 default-tier endpoints for meta-llama/llama-3.3-70b-instruct: DeepInfra (fp8, endpoint deepinfra/turbo), lowest on both input and output as listed.
  • Semantics reviewed 2026-10-15 (#515): OpenRouter's provider-routing guide (https://openrouter.ai/docs/guides/routing/provider-selection.md, Markdown SHA-256 1206e349a01dd137cfc18683eacf7a82bab44c75e5d65d88f0a86b2f212c6e47) states that its default routing load-balances across the lowest-cost stable providers weighted by inverse square of price, and that provider.sort = "price" tries providers in price order; its service-tiers guide (https://openrouter.ai/docs/guides/features/service-tiers.md, Markdown SHA-256 76af5150365778a4775a4218de182b40665fbfd8c8b6babbd623e493b3d84a51) states that flex, priority and ultrafast tier endpoints (tier-suffixed tags) stay opt-in and that provider.sort does not opt into any tier. Its OpenAPI spec (https://openrouter.ai/docs/openapi/openapi.yaml, SHA-256 f36da4e53cd2f259791d0a2110bd1c3b4dd6fb3a1701bc9c6375e79c7df88ed0) defines per-endpoint USD-per-token prices and a discount fraction, and its provider integration guide (https://openrouter.ai/docs/guides/providers/integration.md, Markdown SHA-256 535c7594d2d5197584edf03fb8e4abb4afc64d7beccef929fdf9ed15f58c2224) states that a provider discount is applied to the displayed price users see and pay (user price = base price × (1 - discount_to_user)).
  • Endpoint status codes are undocumented and not used; no availability is asserted. HF IDs and descriptions do not establish checkpoint, developer, license or weight access; no such joins are published.
  • Context window: 131072 tokens, the context_length OpenRouter lists for that endpoint. OpenRouter's provider integration guide (https://openrouter.ai/docs/guides/providers/integration.md, Markdown SHA-256 5881020f822b599ece83ff232645dc31c6ac2d1bd7e354473ebc425951b95742, fetched 2026-10-26T19:37:54Z) defines a provider's declared max_context_length (text) as "the total context window: input and output tokens combined", which "Becomes the endpoint's context length" ("Held if set manually": OpenRouter staff may set it by hand). The value follows the lowest-price endpoint; the model's developer and other endpoints may state different limits.

Raw API evidence ↗ · vendor-api · Claims: openrouter-llama-3-3-70b-instruct: pricing · openrouter-llama-3-3-70b-instruct: specifications · Retrieved · Effective: Unknown · Checked: 2026-11-02T04:30:44.024Z · Reviewed: Unknown · Review after: 2026-11-12T04:30:44.024Z · Freshness: stale · Confidence: B

  • Raw SHA-256: f46452eec1e739ae02229bcc5a67e155a2e39b9e6f2dd497cc2a2304d7fa809b; collector openrouter-endpoints-1; daily collection independently replayed with zero source requests.
  • Lowest listed endpoint of 10 default-tier endpoints for meta-llama/llama-3.3-70b-instruct: DeepInfra (fp8, endpoint deepinfra/turbo), lowest on both input and output as listed.
  • Semantics reviewed 2026-10-15 (#515): OpenRouter's provider-routing guide (https://openrouter.ai/docs/guides/routing/provider-selection.md, Markdown SHA-256 1206e349a01dd137cfc18683eacf7a82bab44c75e5d65d88f0a86b2f212c6e47) states that its default routing load-balances across the lowest-cost stable providers weighted by inverse square of price, and that provider.sort = "price" tries providers in price order; its service-tiers guide (https://openrouter.ai/docs/guides/features/service-tiers.md, Markdown SHA-256 76af5150365778a4775a4218de182b40665fbfd8c8b6babbd623e493b3d84a51) states that flex, priority and ultrafast tier endpoints (tier-suffixed tags) stay opt-in and that provider.sort does not opt into any tier. Its OpenAPI spec (https://openrouter.ai/docs/openapi/openapi.yaml, SHA-256 f36da4e53cd2f259791d0a2110bd1c3b4dd6fb3a1701bc9c6375e79c7df88ed0) defines per-endpoint USD-per-token prices and a discount fraction, and its provider integration guide (https://openrouter.ai/docs/guides/providers/integration.md, Markdown SHA-256 535c7594d2d5197584edf03fb8e4abb4afc64d7beccef929fdf9ed15f58c2224) states that a provider discount is applied to the displayed price users see and pay (user price = base price × (1 - discount_to_user)).
  • Endpoint status codes are undocumented and not used; no availability is asserted. HF IDs and descriptions do not establish checkpoint, developer, license or weight access; no such joins are published.
  • Context window: 131072 tokens, the context_length OpenRouter lists for that endpoint. OpenRouter's provider integration guide (https://openrouter.ai/docs/guides/providers/integration.md, Markdown SHA-256 5881020f822b599ece83ff232645dc31c6ac2d1bd7e354473ebc425951b95742, fetched 2026-10-26T19:37:54Z) defines a provider's declared max_context_length (text) as "the total context window: input and output tokens combined", which "Becomes the endpoint's context length" ("Held if set manually": OpenRouter staff may set it by hand). The value follows the lowest-price endpoint; the model's developer and other endpoints may state different limits.

Raw API evidence ↗ · vendor-api · Claims: openrouter-llama-3-3-70b-instruct: pricing · openrouter-llama-3-3-70b-instruct: specifications · Retrieved · Effective: Unknown · Checked: 2026-11-03T04:29:37.203Z · Reviewed: Unknown · Review after: 2026-11-13T04:29:37.203Z · Freshness: fresh · Confidence: B

  • Raw SHA-256: 55f5d3d39fc4b219c5d2534aaca2bde8e61f450ad505115504a1066d9757319e; collector openrouter-endpoints-1; daily collection independently replayed with zero source requests.
  • Lowest listed endpoint of 10 default-tier endpoints for meta-llama/llama-3.3-70b-instruct: DeepInfra (fp8, endpoint deepinfra/turbo), lowest on both input and output as listed.
  • Semantics reviewed 2026-10-15 (#515): OpenRouter's provider-routing guide (https://openrouter.ai/docs/guides/routing/provider-selection.md, Markdown SHA-256 1206e349a01dd137cfc18683eacf7a82bab44c75e5d65d88f0a86b2f212c6e47) states that its default routing load-balances across the lowest-cost stable providers weighted by inverse square of price, and that provider.sort = "price" tries providers in price order; its service-tiers guide (https://openrouter.ai/docs/guides/features/service-tiers.md, Markdown SHA-256 76af5150365778a4775a4218de182b40665fbfd8c8b6babbd623e493b3d84a51) states that flex, priority and ultrafast tier endpoints (tier-suffixed tags) stay opt-in and that provider.sort does not opt into any tier. Its OpenAPI spec (https://openrouter.ai/docs/openapi/openapi.yaml, SHA-256 f36da4e53cd2f259791d0a2110bd1c3b4dd6fb3a1701bc9c6375e79c7df88ed0) defines per-endpoint USD-per-token prices and a discount fraction, and its provider integration guide (https://openrouter.ai/docs/guides/providers/integration.md, Markdown SHA-256 535c7594d2d5197584edf03fb8e4abb4afc64d7beccef929fdf9ed15f58c2224) states that a provider discount is applied to the displayed price users see and pay (user price = base price × (1 - discount_to_user)).
  • Endpoint status codes are undocumented and not used; no availability is asserted. HF IDs and descriptions do not establish checkpoint, developer, license or weight access; no such joins are published.
  • Context window: 131072 tokens, the context_length OpenRouter lists for that endpoint. OpenRouter's provider integration guide (https://openrouter.ai/docs/guides/providers/integration.md, Markdown SHA-256 5881020f822b599ece83ff232645dc31c6ac2d1bd7e354473ebc425951b95742, fetched 2026-10-26T19:37:54Z) defines a provider's declared max_context_length (text) as "the total context window: input and output tokens combined", which "Becomes the endpoint's context length" ("Held if set manually": OpenRouter staff may set it by hand). The value follows the lowest-price endpoint; the model's developer and other endpoints may state different limits.
About capability scores (Epoch AI)

Capability scores: Epoch AI, 'Capabilities & benchmarking'. Published online at epoch.ai. Retrieved from 'https://epoch.ai/benchmarks' [online resource], epoch.ai/benchmarks. Accessed 2026-10-31. Licensed CC BY 4.0; values here are rounded to one decimal place and grouped into bands.

The Epoch Capabilities Index (ECI) combines many benchmarks on a scale where Claude 3.5 Sonnet is 130 and GPT-5 is 150. Bands, on the displayed one-decimal score: Frontier 160+, Strong 150–159.9, Capable 140–149.9, Entry below 140. Each score has a 90% range; a range that crosses a band boundary is flagged. Models are not ranked: neighbouring scores often overlap.

A benchmark is not your workload. Epoch's scores mostly reflect runs at the highest reasoning effort, which use more output tokens than default settings. Scores are dated and can shift between Epoch releases. A score belongs to the developer's model; a host listing shows the same model's score, but hosts may serve different versions or quantizations.

How headline rates are chosen

Headline rates use the directory's shared rule: the rate a buyer meets by default, chosen by billing, tier and context dimensions and never by price. Context caps, the bracket starting at zero and short-context variants are labeled; unverified availability is flagged and is not a service guarantee. Batch, off-peak and other variants remain in the details; ambiguous headlines stay unknown.

No currency conversion is performed. Missing cache prices are unknown, not zero. Equal units do not establish equivalent models or service conditions. See the methodology and source register.