API

JSON endpoints for LLM API prices, offers and price history.

Free
No key needed
JSON
CORS open

$ curl "https://api.inferindex.dev/cheapest?model=deepseek/deepseek-v3.2"

cheapest.provider: GMICloud
cheapest.input_per_1M: 0.2088
cheapest.blended_per_1M: 0.234

Response right now · HTTP 200See the response

Response right now · HTTP 200

{
  "model": {
    "id": "deepseek/deepseek-v3.2",
    "name": "DeepSeek: DeepSeek V3.2",
    "new_listing": null,
    "retirement": null
  },
  "cheapest": {
    "provider": "GMICloud",
    "via": "openrouter",
    "input_per_1M": 0.2088,
    "output_per_1M": 0.3096,
    "blended_per_1M": 0.234,
    "currency": "USD",
    "quantization": "fp8",
    "context_length": 163840,
    "promo": true,
    "price_before_promo": {
      "input_per_1M": 0.29,
      "output_per_1M": 0.43
    },
    "checked_at": "2026-10-31T08:50:41.527Z",
    "stale": false
  },
  "offers_total": 21,
  "fetched_at": "2026-10-31T10:05:51.479Z"
}

Returned by the API while this page was generated, trimmed to the main fields.

Start here

Find the model id/models?search=deepseekreturns ids such as deepseek/deepseek-v3.2.
Ask for the cheapest offer/cheapest?model=deepseek/deepseek-v3.2one request, no key.
Read the answercheapest.provider, cheapest.via and cheapest.blended_per_1Mthe provider serving the model, where you buy it (direct or through an aggregator), and its blended price in USD per million tokens.

Rate limit · per minute

Without a key: 60
With a free key: 300

The API works without a key, at 60 requests a minute. A free key raises that to 300. Get a free key

Examples

What does a model cost right now, and who sells it?

curl

curl "https://api.inferindex.dev/cheapest?model=meta-llama/llama-3.3-70b-instruct&limit=3"

Response as of 2026-10-16, abridged.

JSON

{
  "model": { "id": "meta-llama/llama-3.3-70b-instruct", "name": "Meta: Llama 3.3 70B Instruct" },
  "offers_total": 41,
  "cheapest": {
    "provider": "NanoGPT",
    "via": "nanogpt",
    "via_name": "NanoGPT",
    "input_per_1M": 0.05,
    "output_per_1M": 0.23,
    "blended_per_1M": 0.095,
    "promo": false,
    "checked_at": "2026-10-16T07:40:37.737Z"
  }
}

Read: blended_per_1M is the price at a 3:1 input:output mix, and via_name the seller when the offer is sold through an aggregator. promo flags a drop or a promotion (false here; probable is a drop we saw, not announced by the seller) and checked_at is when the price was last confirmed.

What would my workload cost per month?

curl

curl "https://api.inferindex.dev/cheapest?model=openai/gpt-oss-120b&prompt_tokens=2000&output_tokens=500&requests_per_day=1000&limit=3"

Response as of 2026-10-16, abridged.

JSON

{
  "model": { "id": "openai/gpt-oss-120b", "name": "OpenAI: gpt-oss-120b" },
  "usage": { "prompt_tokens": 2000, "output_tokens": 500, "cached_ratio": 0, "requests_per_day": 1000 },
  "cheapest": {
    "provider": "Runware",
    "via_name": "LLM Gateway",
    "input_per_1M": 0.032,
    "output_per_1M": 0.14,
    "estimated_cost_per_request": 0.000134,
    "estimated_cost_per_month": 4.02,
    "estimate_approx": false,
    "estimate_notes": ["context_unknown", "max_output_unknown"]
  }
}

Read: estimated_cost_per_month covers 30 days at the stated workload, for the offer that is cheapest for that workload; estimate_approx is true when the source does not fully publish its tiers or cache price; estimate_notes lists caveats such as context_unknown.

Which of two models is cheaper for the same workload?

There is no route that compares two models. Call /cheapest once per model with the same workload and compare estimated_cost_per_month:

curl

curl "https://api.inferindex.dev/cheapest?model=meta-llama/llama-3.3-70b-instruct&prompt_tokens=2000&output_tokens=500&requests_per_day=1000&limit=1"
curl "https://api.inferindex.dev/cheapest?model=openai/gpt-oss-120b&prompt_tokens=2000&output_tokens=500&requests_per_day=1000&limit=1"

Responses as of 2026-10-16, abridged.

JSON

{ "model": { "id": "meta-llama/llama-3.3-70b-instruct" }, "cheapest": { "provider": "NanoGPT", "via_name": "NanoGPT", "estimated_cost_per_month": 6.45 } }
{ "model": { "id": "openai/gpt-oss-120b" },    "cheapest": { "provider": "Runware",  "via_name": "LLM Gateway", "estimated_cost_per_month": 4.02 } }

Read: the two figures are comparable because the workload is the same; the website's comparison page shows them side by side. No route lists every offer of one provider either: the provider pages of the website do.

Has the price changed?

curl

curl "https://api.inferindex.dev/history?model=deepseek/deepseek-v3.2&days=7&granularity=day&provider=DeepInfra"

Response as of 2026-10-16, first two points of sixteen.

JSON

{
  "model": { "id": "deepseek/deepseek-v3.2" },
  "days": 7,
  "granularity": "day",
  "history": [
    { "period": "2026-10-09", "provider": "DeepInfra", "quantization": "fp4", "variant": "", "input_min": 0.26, "input_max": 0.26, "input_last": 0.26, "output_min": 0.38, "output_max": 0.38, "output_last": 0.38 },
    { "period": "2026-10-09", "provider": "DeepInfra", "quantization": "fp4", "variant": "flex", "input_min": 0.208, "input_max": 0.208, "input_last": 0.208, "output_min": 0.304, "output_max": 0.304, "output_last": 0.304 }
  ]
}

Read: one point per day and per offer; equal min, max and last mean the price did not move that day. variant: "flex" is a lower-priority tier, not the standard price.

Who offers it in the EU, without training on my prompts?

curl

curl "https://api.inferindex.dev/resellers?model=meta-llama/llama-3.3-70b-instruct&region=eu&no_training=true&strict=true&limit=3"

Response as of 2026-10-16, abridged.

JSON

{
  "model": { "id": "meta-llama/llama-3.3-70b-instruct" },
  "total_offers": 4,
  "filters_unknown": { "region": 54, "no_training": 26, "quantization": 9 },
  "excluded": { "region_unknown_strict": 54, "region_mismatch": 8, "training_unknown_strict": 26, "training_not_excluded": 6 },
  "offers": [
    { "provider": "CloudFerro", "blended_per_1M": 0.4763, "conditions": { "provider": { "regions": { "value": "EU (Poland)" }, "training_on_prompts": { "value": "no" } } } },
    { "provider": "IONOS",      "blended_per_1M": 0.7285, "conditions": { "provider": { "regions": { "value": "Germany (EU)" }, "training_on_prompts": { "value": "no" } } } }
  ]
}

Read: without strict=true, offers whose region or policy is not published are kept and counted in filters_unknown; "not published" is not "no". With strict=true they are excluded, which is why the list is short; excluded says how many fell out and why.

Rent a GPU or pay the API?

curl

curl "https://api.inferindex.dev/self-host?model=qwen/qwen3-32b&tokens_per_day=50000000"

Response as of 2026-10-16, abridged.

JSON

{
  "status": "ok",
  "verdict": "api_cheaper",
  "headline": "At 50 million tokens a day, the API is cheaper whatever the throughput ($0.123/M at a 3:1 input:output mix; self-hosting breaks even at 68.6 million tokens a day).",
  "threshold_m_tokens_per_day": 68.57,
  "required_tokens_per_second": { "at_utilization": 1587.4, "utilization_percent": 50 },
  "api": { "provider": "Nscale", "usd_per_m_tokens": 0.1225, "offers_compared": 21 },
  "assumptions": ["Hardware rental only: the hourly price the provider publishes for the machine. It does not count the engineering to deploy and run the model, monitoring, …", "…"],
  "configuration": { "gpu_name": "RTX A6000 48GB", "gpus_needed": 1, "provider": "Thunder Compute", "usd_per_day": 8.4, "throughput_evidence": "none" }
}

Read: threshold_m_tokens_per_day is the volume at which a rented configuration costs the same as the cheapest API; above it, renting wins only if the setup sustains required_tokens_per_second, a condition and not a promise. The figures cover hardware rental only; assumptions lists what is left out.

Where does the market stand, by model family?

curl

curl "https://api.inferindex.dev/index?series=market"

Response for the publication of 2026-10-12, abridged.

JSON

{
  "index": "InferIndex market index",
  "series": "market",
  "published_at": "2026-10-12T00:00:00.000Z",
  "method_version": "1.2",
  "unit": "USD per 1M tokens, blended 3:1 (input:output)",
  "compared_to": "2026-10-05T00:00:00.000Z",
  "families": [
    { "family": "open_small", "value": 0.125, "models": 54, "change_on_same_models_pct": 0, "same_models": 54 },
    { "family": "open_medium", "value": 0.2953, "models": 30, "change_on_same_models_pct": 2.16, "same_models": 30 }
  ],
  "corrections": []
}

Read: each family's value is the median blended price of the models counted; change_on_same_models_pct compares only the models counted in both publications, so a model entering or leaving does not move it. corrections lists any value corrected after publication, with the reason (empty here).

Where does a model stand in the public leaderboards?

curl

curl "https://api.inferindex.dev/rankings?model=deepseek/deepseek-v3.2"

Response as of 2026-10-16, abridged.

JSON

{
  "model": { "id": "deepseek/deepseek-v3.2" },
  "rankings": [
    { "source": "Arena", "bench": "text_style_control", "listed_as": "deepseek-v3.2", "rank": 123, "out_of": 415, "score": 1424.67, "votes": 48134, "published_on": "2026-10-13",
      "scale": { "score_name": "Arena Score", "count_name": "votes" } }
  ],
  "attribution": { "license": "CC BY 4.0", "license_url": "https://creativecommons.org/licenses/by/4.0/" }
}

Read: the rank and score are the source's, never computed by us; scale says what the score is called and what the count is. Keep the attribution object when you reuse the data: it is licensed under CC BY 4.0.

Endpoints

GET/cheapestThe cheapest current offer for a model, with its conditions and why it was chosen.

model: Model id or name (required)
no_training: true: exclude offers that may train on prompts (unpublished policy kept and flagged unless strict=true)
region: eu, us…: processing region of provider and aggregator; unknown region kept unless strict=true

7 more parameters

quantization: fp8, fp4…: precision
min_context: Minimum context length
tools, json, vision: true: capability declared by the source
strict: true: exclude offers missing the filtered data
prompt_tokens, output_tokens, requests_per_day: Estimated cost per request and per month
offers, limit: offers=false: winner only; limit: first N offers
More: sort, include_tiers, include_stale, explain, min_uptime, no_waitlist, cached_ratio: see the full list on the API root

GET/resellersEvery current offer for a model, direct and via aggregators, one line per offer (provider, source, precision, tier).

model: Model id or name (required)
region, no_training, no_waitlist, strict: Same conditions as /cheapest
include_tiers: flex, batch: show lower-priority tiers

1 more parameter

sort, limit, cursor: Order and pagination

Request

$ curl "https://api.inferindex.dev/resellers?model=deepseek/deepseek-v3.2&limit=2"

Response right now · HTTP 200

{
  "model": {
    "id": "deepseek/deepseek-v3.2",
    "name": "DeepSeek: DeepSeek V3.2",
    "new_listing": null,
    "retirement": null
  },
  "total_offers": 53,
  "offers": [
    {
      "provider": "NanoGPT",
      "via": "nanogpt",
      "via_name": "NanoGPT",
      "blended_per_1M": 0.234,
      "promo": "probable",
      "stale": false
    },
    {
      "provider": "GMICloud",
      "via": "openrouter",
      "via_name": "OpenRouter",
      "blended_per_1M": 0.234,
      "promo": true,
      "stale": false
    }
  ]
}

Returned by the API while this page was generated, trimmed to the main fields.

GET/historyPrice history: the cheapest offer of each day, every offer day by day, or the prices in force at a date, plus the lab's official prices back to 2023.

model: Model id or name (required)
series=cheapest: The offer /cheapest would have selected at 00:00 UTC each day
days, from, to, at: Period, or a single instant

2 more parameters

granularity: raw, day, week
provider, limit, cursor: Filter and pagination

Request

$ curl "https://api.inferindex.dev/history?model=deepseek/deepseek-v3.2&series=cheapest&days=7"

Response right now · HTTP 200

{
  "model": {
    "id": "deepseek/deepseek-v3.2",
    "name": "DeepSeek: DeepSeek V3.2",
    "new_listing": null
  },
  "series": "cheapest",
  "count": 7,
  "cheapest_by_day": [
    {
      "date": "2026-10-30",
      "at": "2026-10-30T00:00:00.000Z",
      "cheapest": {
        "provider": "GMICloud",
        "via": "openrouter",
        "variant": null,
        "quantization": "fp8",
        "input_per_1M": 0.2088,
        "output_per_1M": 0.3096,
        "blended_per_1M": 0.234,
        "promo": true,
        "platform_fee_rate": 0.055
      },
      "providers": 22
    },
    {
      "date": "2026-10-31",
      "at": "2026-10-31T00:00:00.000Z",
      "cheapest": {
        "provider": "GMICloud",
        "via": "openrouter",
        "variant": null,
        "quantization": "fp8",
        "input_per_1M": 0.2088,
        "output_per_1M": 0.3096,
        "blended_per_1M": 0.234,
        "promo": true,
        "platform_fee_rate": 0.055
      },
      "providers": 22
    }
  ]
}

Returned by the API while this page was generated, trimmed to the main fields.

GET/modelsSearch the model catalogue.

search: Model or lab name

Request

$ curl "https://api.inferindex.dev/models?search=deepseek"

Response right now · HTTP 200

{
  "count": 20,
  "models": [
    {
      "id": "deepseek/deepseek-v4.1-flash-1017",
      "name": "DeepSeek: DeepSeek V4.1 Flash 1017"
    },
    {
      "id": "deepseek/deepseek-v4.1-flash",
      "name": "DeepSeek: DeepSeek V4.1 Flash"
    },
    {
      "id": "deepseek/deepseek-v4-flash-vision-exp",
      "name": "DeepSeek: DeepSeek V4 Flash Vision Exp"
    }
  ]
}

Returned by the API while this page was generated, trimmed to the main fields.

GET/indexThe weekly market index: median price per model family, published every Monday.

series: market (offers we collect) or official (labs' list prices)
at: YYYY-MM-DD: the latest publication on or before that date

Request

$ curl "https://api.inferindex.dev/index?series=market"

Response right now · HTTP 200

{
  "index": "InferIndex market index",
  "series": "market",
  "published_at": "2026-10-26T00:00:00.000Z",
  "method_version": "1.3",
  "unit": "USD per 1M tokens, blended 3:1 (input:output)",
  "compared_to": "2026-10-19T00:00:00.000Z",
  "families": [
    {
      "family": "frontier_closed",
      "value": 3.5,
      "models": 95,
      "model_ids": [
        "aion-labs/aion-2.0",
        "aion-labs/aion-3.0",
        "aion-labs/aion-3.5",
        "amazon/nova-premier-v1",
        "amazon/nova-pro-v1",
        "anthropic/claude-fable-5",
        "anthropic/claude-fable-5.1",
        "anthropic/claude-opus-4.5",
        "anthropic/claude-opus-4.6",
        "anthropic/claude-opus-4.7",
        "anthropic/claude-opus-4.8",
        "anthropic/claude-opus-5",
        "anthropic/claude-opus-5.5",
        "anthropic/claude-sonnet-4.5",
        "anthropic/claude-sonnet-4.6",
        "anthropic/claude-sonnet-5",
        "anthropic/claude-sonnet-5.5",
        "bytedance-seed/doubao-seed-2.1-pro",
        "bytedance-seed/doubao-seed-2.1-turbo",
        "bytedance-seed/seed-1.6",
        "bytedance-seed/seed-2-1-turbo",
        "google/gemini-2.5-pro",
        "google/gemini-2.5-pro-preview",
        "google/gemini-3.1-pro-preview",
        "google/gemini-3.1-pro-preview-customtools",
        "inception/mercury-2",
        "inception/mercury-3",
        "meta/muse-spark-1.1",
        "meta/muse-spark-1.2",
        "meta/muse-spark-1.2-contributor",
        "meta/muse-spark-1.3",
        "meta/muse-spark-1.3-contributor",
        "meta/muse-spark-open",
        "minimax/minimax-m2-her",
        "mistralai/mistral-medium-3",
        "mistralai/mistral-medium-3.1",
        "mistralai/mistral-saba",
        "openai/gpt-3.5-turbo-16k",
        "openai/gpt-4.1",
        "openai/gpt-4o",
        "openai/gpt-4o-2024-08-06",
        "openai/gpt-4o-2024-11-20",
        "openai/gpt-5",
        "openai/gpt-5-pro",
        "openai/gpt-5.1",
        "openai/gpt-5.2",
        "openai/gpt-5.2-chat",
        "openai/gpt-5.2-pro",
        "openai/gpt-5.4",
        "openai/gpt-5.4-pro",
        "openai/gpt-5.5",
        "openai/gpt-5.5-pro",
        "openai/gpt-5.6-sol",
        "openai/gpt-5.6-sol-pro",
        "openai/gpt-5.6-terra",
        "openai/gpt-5.6-terra-pro",
        "openai/gpt-6-astra",
        "openai/gpt-6-astra-pro",
        "openai/gpt-6-sol",
        "openai/gpt-6-sol-pro",
        "openai/gpt-6.1-sol",
        "openai/gpt-6.1-sol-pro",
        "openai/gpt-chat-latest",
        "openai/o3",
        "openai/o3-pro",
        "perceptron/perceptron-mk1",
        "perceptron/perceptron-mk1.5",
        "perplexity/sonar-deep-research",
        "perplexity/sonar-pro",
        "perplexity/sonar-reasoning-pro",
        "qwen/qwen-plus",
        "qwen/qwen3-max",
        "qwen/qwen3-max-thinking",
        "qwen/qwen3-vl-plus",
        "qwen/qwen3.5-plus",
        "qwen/qwen3.6-plus",
        "qwen/qwen3.7-max",
        "qwen/qwen3.7-plus",
        "qwen/qwen3.8-max",
        "qwen/qwen3.8-max-0902",
        "qwen/qwen3.8-max-prime",
        "sakana/fugu-max",
        "sakana/fugu-ultra",
        "sakana/fugu-ultra-v2",
        "upstage/solar-pro-3",
        "upstage/solar-pro4",
        "writer/palmyra-x5",
        "x-ai/grok-4.20",
        "x-ai/grok-4.20-multi-agent",
        "x-ai/grok-4.3",
        "x-ai/grok-4.5",
        "x-ai/grok-4.6",
        "x-ai/grok-4.7",
        "x-ai/grok-build-0.1",
        "z-ai/glm-5v-turbo"
      ],
      "change_on_same_models_pct": 0,
      "same_models": 94,
      "models_entered": [
        "inception/mercury-3"
      ],
      "models_left": [
        "anthropic/claude-opus-4",
        "anthropic/claude-opus-4.1",
        "anthropic/claude-sonnet-4",
        "openai/gpt-3.5-turbo",
        "openai/gpt-4",
        "openai/gpt-4-turbo",
        "openai/gpt-4o-2024-05-13",
        "openai/o1",
        "openai/o1-pro"
      ],
      "excluded_from_change": []
    },
    {
      "family": "compact_closed",
      "value": 0.4625,
      "models": 37,
      "model_ids": [
        "aion-labs/aion-3.0-mini",
        "aion-labs/aion-3.5-mini",
        "amazon/nova-2-lite-v1",
        "amazon/nova-lite-v1",
        "amazon/nova-micro-v1",
        "anthropic/claude-haiku-4.5",
        "anthropic/claude-haiku-5.5",
        "bytedance-seed/seed-1.6-flash",
        "bytedance-seed/seed-2.0-lite",
        "bytedance-seed/seed-2.0-mini",
        "google/gemini-2.5-flash",
        "google/gemini-2.5-flash-lite",
        "google/gemini-3-flash-preview",
        "google/gemini-3.1-flash-lite",
        "google/gemini-3.1-flash-lite-preview",
        "google/gemini-3.5-flash-lite",
        "google/gemini-3.6-flash",
        "google/gemini-3.8-flash",
        "inception/mercury-2.5",
        "openai/gpt-4.1-mini",
        "openai/gpt-4o-mini",
        "openai/gpt-4o-mini-2024-07-18",
        "openai/gpt-5-mini",
        "openai/gpt-5-nano",
        "openai/gpt-5.4-mini",
        "openai/gpt-5.4-nano",
        "openai/gpt-5.6-luna",
        "openai/gpt-5.6-luna-pro",
        "openai/gpt-6-luna",
        "openai/gpt-6-luna-pro",
        "openai/o3-mini-high",
        "openai/o4-mini-high",
        "perplexity/sonar",
        "qwen/qwen3.5-flash",
        "qwen/qwen3.6-flash",
        "qwen/qwen3.7-flash",
        "z-ai/glm-5.3-flashx"
      ],
      "change_on_same_models_pct": 0,
      "same_models": 37,
      "models_entered": [],
      "models_left": [
        "openai/gpt-4.1-nano",
        "openai/o3-mini",
        "openai/o4-mini"
      ],
      "excluded_from_change": []
    },
    {
      "family": "code",
      "value": 0.4625,
      "models": 10,
      "model_ids": [
        "bytedance-seed/seed-2.0-code",
        "mistralai/codestral-2508",
        "mistralai/devstral-2512",
        "moonshotai/kimi-k2.7-code",
        "openai/gpt-5.3-codex",
        "qwen/qwen-2.5-coder-32b-instruct",
        "qwen/qwen3-coder",
        "qwen/qwen3-coder-30b-a3b-instruct",
        "qwen/qwen3-coder-flash",
        "qwen/qwen3-coder-next"
      ],
      "change_on_same_models_pct": 0,
      "same_models": 10,
      "models_entered": [],
      "models_left": [
        "openai/gpt-5.1-codex",
        "openai/gpt-5.1-codex-max",
        "openai/gpt-5.1-codex-mini",
        "openai/gpt-5.2-codex"
      ],
      "excluded_from_change": []
    },
    {
      "family": "open_large",
      "value": 0.5,
      "models": 61,
      "model_ids": [
        "arcee-ai/trinity-large-thinking",
        "deepseek/deepseek-chat",
        "deepseek/deepseek-chat-v3-0324",
        "deepseek/deepseek-chat-v3.1",
        "deepseek/deepseek-r1",
        "deepseek/deepseek-r1-0528",
        "deepseek/deepseek-v3.1-terminus",
        "deepseek/deepseek-v3.2",
        "deepseek/deepseek-v3.2-exp",
        "deepseek/deepseek-v3.2-exp-thinking",
        "deepseek/deepseek-v4-flash",
        "deepseek/deepseek-v4-flash-0731",
        "deepseek/deepseek-v4-flash-vision-exp",
        "deepseek/deepseek-v4-pro",
        "deepseek/deepseek-v4-pro-0813",
        "deepseek/deepseek-v4.1-flash",
        "inclusionai/ling-3.1-flash",
        "meituan/longcat-2.0",
        "meta-llama/llama-4-maverick",
        "minimax/minimax-01",
        "minimax/minimax-m1",
        "minimax/minimax-m2",
        "minimax/minimax-m2.1",
        "minimax/minimax-m2.5",
        "minimax/minimax-m2.7",
        "minimax/minimax-m3",
        "minimaxai/minimax-m1-80k",
        "moonshotai/kimi-k2",
        "moonshotai/kimi-k2-0905",
        "moonshotai/kimi-k2-thinking",
        "moonshotai/kimi-k2.5",
        "moonshotai/kimi-k2.6",
        "moonshotai/kimi-k3",
        "nousresearch/hermes-3-llama-3.1-405b",
        "nousresearch/hermes-4-405b",
        "nvidia/nemotron-3-ultra-550b-a55b",
        "qwen/qwen3-235b-a22b",
        "qwen/qwen3-235b-a22b-2507",
        "qwen/qwen3-235b-a22b-thinking-2507",
        "qwen/qwen3-vl-235b-a22b-instruct",
        "qwen/qwen3-vl-235b-a22b-thinking",
        "qwen/qwen3.5-397b-a17b",
        "qwen/qwen3.8-2.4t-a95b",
        "stepfun/step-3.7-flash",
        "stepfun/step-5-preview",
        "tencent/hy3",
        "tencent/hy3-preview",
        "tencent/hy4-preview",
        "thinkingmachines/inkling",
        "thinkingmachines/inkling-small",
        "xiaomi/mimo-v2.5",
        "xiaomi/mimo-v2.5-pro",
        "xiaomi/mimo-v2.6-pro",
        "z-ai/glm-4.5",
        "z-ai/glm-4.6",
        "z-ai/glm-4.7",
        "z-ai/glm-5",
        "z-ai/glm-5.1",
        "z-ai/glm-5.2",
        "z-ai/glm-5.3",
        "z-ai/glm-5.3-flash"
      ],
      "change_on_same_models_pct": 11.06,
      "same_models": 61,
      "models_entered": [],
      "models_left": [],
      "excluded_from_change": []
    },
    {
      "family": "open_medium",
      "value": 0.329,
      "models": 30,
      "model_ids": [
        "anthracite-org/magnum-v4-72b",
        "cohere/command-a",
        "cohere/command-r-plus-08-2024",
        "deepseek/deepseek-r1-distill-llama-70b",
        "inclusionai/ling-3.0-flash",
        "inclusionai/ling-3.0-flash-fin",
        "inclusionai/ling-3.0-flash-vl",
        "meta-llama/llama-3.1-70b-instruct",
        "meta-llama/llama-3.3-70b-instruct",
        "meta-llama/llama-4-scout",
        "mistralai/mistral-large-2407",
        "mistralai/mistral-medium-3-5",
        "mistralai/mistral-small-2603",
        "mistralai/mixtral-8x22b-instruct",
        "nousresearch/hermes-3-llama-3.1-70b",
        "nvidia/nemotron-3-super-120b-a12b",
        "openai/gpt-oss-120b",
        "poolside/laguna-s-2.1",
        "qwen/qwen-2.5-72b-instruct",
        "qwen/qwen2.5-vl-72b-instruct",
        "qwen/qwen3-next-80b-a3b-instruct",
        "qwen/qwen3-next-80b-a3b-thinking",
        "qwen/qwen3.5-122b-a10b",
        "qwen/qwen3.8-flash",
        "sao10k/l3.3-euryale-70b",
        "stepfun/step-3.5-flash",
        "xiaomi/mimo-v2.6-flash",
        "z-ai/glm-4.5-air",
        "z-ai/glm-4.5v",
        "z-ai/glm-4.6v"
      ],
      "change_on_same_models_pct": 0,
      "same_models": 30,
      "models_entered": [],
      "models_left": [],
      "excluded_from_change": []
    },
    {
      "family": "open_small",
      "value": 0.1275,
      "models": 56,
      "model_ids": [
        "aion-labs/aion-rp-llama-3.1-8b",
        "aisingapore/gemma-sea-lion-v4-27b-it",
        "aisingapore/qwen-sea-lion-v4-32b-it",
        "bytedance/ui-tars-1.5-7b",
        "cohere/command-r-08-2024",
        "cohere/command-r7b-12-2024",
        "google/gemma-2-27b-it",
        "google/gemma-3-12b-it",
        "google/gemma-3-27b-it",
        "google/gemma-3-4b-it",
        "google/gemma-4-26b-a4b-it",
        "google/gemma-4-31b-it",
        "ibm-granite/granite-4.0-h-micro",
        "ibm-granite/granite-4.2-8b",
        "liquid/lfm-2.5-24b-a2b",
        "meta-llama/llama-3.1-8b-instruct",
        "meta-llama/llama-3.2-1b-instruct",
        "meta-llama/llama-3.2-3b-instruct",
        "meta/muse-glimmer-30b",
        "microsoft/phi-4",
        "mistralai/ministral-14b-2512",
        "mistralai/ministral-3b-2512",
        "mistralai/ministral-8b-2512",
        "mistralai/mistral-nemo",
        "mistralai/mistral-small-24b-instruct-2501",
        "mistralai/mistral-small-3.1-24b-instruct",
        "mistralai/mistral-small-3.2-24b-instruct",
        "mistralai/voxtral-small-24b-2507",
        "nex-agi/nex-n2.5-mini",
        "nvidia/nemotron-3-nano-30b-a3b",
        "nvidia/nemotron-3.5-lightning",
        "openai/gpt-oss-20b",
        "poolside/laguna-xs-2.1",
        "qwen/qwen-2.5-7b-instruct",
        "qwen/qwen3-14b",
        "qwen/qwen3-30b-a3b",
        "qwen/qwen3-30b-a3b-instruct-2507",
        "qwen/qwen3-30b-a3b-thinking-2507",
        "qwen/qwen3-32b",
        "qwen/qwen3-8b",
        "qwen/qwen3-vl-30b-a3b-instruct",
        "qwen/qwen3-vl-30b-a3b-thinking",
        "qwen/qwen3-vl-8b-instruct",
        "qwen/qwen3.5-27b",
        "qwen/qwen3.5-35b-a3b",
        "qwen/qwen3.5-9b",
        "qwen/qwen3.6-27b",
        "qwen/qwen3.6-35b-a3b",
        "qwen/qwen3.8-27b",
        "rekaai/reka-edge",
        "sao10k/l3-8b-stheno-v3.2",
        "sao10k/l3-lunaris-8b",
        "speakleash/bielik-11b-v3.0-instruct",
        "thedrummer/unslopnemo-12b",
        "z-ai/glm-4.7-flash",
        "zai-org/glm-4.6v-flash"
      ],
      "change_on_same_models_pct": 0,
      "same_models": 56,
      "models_entered": [],
      "models_left": [],
      "excluded_from_change": []
    }
  ],
  "available_dates": [
    "2026-10-26",
    "2026-10-19",
    "2026-10-12",
    "2026-10-05"
  ],
  "archives": [
    {
      "kind": "publication",
      "url": "https://api.inferindex.dev/index/archive/2026-10-26/publication.json",
      "sha256": "adf67e5ddd08017daa47cce43cf51ea5cbf554a5b017637db9541feb7a8e3a0b",
      "bytes": 18739,
      "archived_at": "2026-10-26T00:04:07.203Z"
    },
    {
      "kind": "correction",
      "url": "https://api.inferindex.dev/index/archive/2026-10-26/correction-2026-10-30T19-35-01-503Z.json",
      "sha256": "4682871037e8d1638ce3c81f34e52f8dd8eead6076e065fed6a073f3fdd9d7ae",
      "bytes": 16015,
      "archived_at": "2026-10-30T19:35:01.503Z"
    }
  ],
  "corrections": []
}

Returned by the API while this page was generated, trimmed to the main fields.

GET/provider-rankingProviders ranked by their median gap to the market price, refreshed every hour.

Request

$ curl "https://api.inferindex.dev/provider-ranking"

Response right now · HTTP 200

{
  "computed_at": "2026-10-31T09:13:08.405Z",
  "min_models": 10,
  "min_providers": 3,
  "providers": [
    {
      "provider": "Wafer",
      "cheapest_count": 0,
      "score_pct": -44.6,
      "models_measured": 10,
      "rank": 1
    },
    {
      "provider": "Relace",
      "cheapest_count": 2,
      "score_pct": -44.1,
      "models_measured": 10,
      "rank": 2
    },
    {
      "provider": "Darkbloom",
      "cheapest_count": 4,
      "score_pct": -43.4,
      "models_measured": 10,
      "rank": 3
    }
  ]
}

Returned by the API while this page was generated, trimmed to the main fields.

Signals on every offer

Full documentation Self-host or API?MCP server

InferIndex only reports prices: it does not route your requests or sell access, and no key is required. An offer with a signal stays listed; some signals keep it from being selected as the cheapest. /cheapest, /resellers, /history, /self-host, /gpus and /gpu-rentals allow 60 uncached requests per minute per IP. /cheapest and /resellers answer at once with the last computed response, up to 30 minutes old, and refresh it in the background. The Age header gives its age in seconds, X-InferIndex-Computed-At the time it was computed, and each offer's checked_at the last time its price was read.