AI Intelligence Calculator

    What's new

    AI model prices change constantly. This page logs notable additions and corrections to the catalog, and lists the most recently released models we track. Every figure on the site carries a last-verified date, and the whole catalog is refreshed weekly from the OpenRouter API.

    Latest changes

    2026-09-19 — Step 5 Preview added — a new lab at the top of the value table

    StepFun's Step 5 Preview joins the catalogue with an index of 44 at $1 / $2.70 per 1M tokens — above Grok 4.6 and Kimi K3, for roughly a tenth of the premium flagships' price. It takes text and image input over a 1M-token context and runs at about 100 tokens per second. Two caveats: it is a preview release, and StepFun publishes no first-party USD price list we could reach, so the price comes from Artificial Analysis. It carries a review flag until we can confirm it.

    2026-09-19 — GLM-5.3 FlashX added

    Z.ai's high-speed variant of GLM-5.3 Flash is in the catalogue at $0.37 / $1.25 per 1M tokens with a 1M-token context window and roughly 200 tokens per second — about 2.5x GLM-5.3 Flash's per-token price for a large jump in throughput. Artificial Analysis has not benchmarked it yet, so it carries no Intelligence Index and its rating is a tier estimate rather than a measurement.

    2026-09-19 — Four superseded models retired

    Claude Fable 5, Grok 4.5, GLM-5.2 and Gemini 3.6 Flash are now marked obsolete, as each has been replaced by a newer model from the same provider that matches or beats it: Fable 5.1 (53 vs 50 at the same $10 / $50), Grok 4.6 (44 vs 39 at the same $2 / $6), GLM-5.3 (45 vs 34 at the same $1.40 / $4.40) and Gemini 3.8 Flash (41 vs 34 at the same $0.75 / $3.75). Their pages stay online with a link to the successor, and they are hidden from the calculator unless you turn retired models on. Gemini 3.7 Flash remains current — Google still sells it and it is only a month old.

    2026-09-19 — Qwen3.8 Max re-benchmarked to 45

    Artificial Analysis has revised Qwen3.8 Max's 0902 snapshot upward, from 40 to 45 on the Intelligence Index v4.3. That places Alibaba's flagship level with GLM-5.3 and above Grok 4.6 at $2 / $6 per 1M tokens, and it is now the strongest multimodal model under $10 per 1M output on the site.

    2026-09-19 — Claude Sonnet 5's price is now permanent

    Anthropic has confirmed that the $2 / $10 per 1M input/output pricing for Claude Sonnet 5 — announced at launch as introductory pricing through 31 August 2026 — is now the standard price, and that the previously scheduled increase to $3 / $15 will not happen. No change was needed on our side; the figure we publish was already the correct one.

    2026-09-13 — Retired models now keep their pages

    When a provider retires a model we no longer delete its page. Retired models stay online with their final verified prices, a clear "Retired model" notice and a link to the model that replaces them — so links and search results for superseded models keep working instead of landing on a 404. We also added a proper 404 page that routes visitors back to the calculator.

    2026-09-13 — New tools: spend planner, per-task costs and plan pages

    Three new ways to use the site: an AI spend planner that adds up every subscription and API budget you pay for and checks whether each plan beats metered pricing; a "what one task costs" table on every model page (a code review, a landing page, an agent run) priced in dollars rather than per-million tokens; and a dedicated page for each of the 14 subscription plans. We also added model-vs-model comparisons such as GPT-6 Astra vs Claude Fable 5.1 and DeepSeek V4.1 Flash vs V4 Pro.

    2026-09-13 — Comparisons re-pointed at the current generation

    Every head-to-head comparison now names the exact models it discusses, so the cards can never drift from the copy. OpenAI comparisons use GPT-6 Astra, Claude comparisons Claude Fable 5.1, Grok comparisons Grok 4.6 and Google comparisons Gemini 3.8 Flash — the models each provider actually leads with today. Stale index figures left over from the previous benchmark scale were corrected as well.

    2026-09-13 — Catalog re-synced to Artificial Analysis Intelligence Index v4.3

    Artificial Analysis re-scaled its Intelligence Index, so every model we track has been re-indexed against the current v4.3 scores. Indices are lower across the board (GPT-5.6 Sol 61 → 47, Claude Opus 5 63 → 51, Gemini 3.7 Flash 56 → 39) but the ranking order is unchanged. Every price was re-verified against the provider's official pricing page and all last-verified dates now read 13 Sep 2026.

    2026-09-13 — Claude Fable 5.1, GPT-6 Astra, Gemini 3.8 Flash and five more models added

    Added Claude Fable 5.1 (index 53, $10 / $50 — the highest-indexed model we track), GPT-6 Astra (53, $10 / $50), Gemini 3.8 Flash (41, $0.75 / $3.75), DeepSeek V4.1 Flash (40, $0.15 / $0.60 off-peak), GLM-5.3 Flash (42, $0.15 / $0.50), Qwen3.8 Max (40, $2 / $6), Meta's Muse Spark 1.3 (48, $1.25 / $4.25) and Sakana AI's Fugu Max and Fugu Ultra v2 multi-agent systems. ChatGPT Plus/Pro and the Claude plans now list the new models.

    2026-09-13 — DeepSeek V4 Flash retired

    DeepSeek's documentation now states that the v4-flash name is retired and that requests are served by DeepSeek V4.1 Flash at the same Flash pricing. V4 Flash is marked obsolete and DeepSeek V4.1 Flash takes its place in the catalogue, in the DeepSeek app plan and in our cost guides.

    2026-08-24 — DeepSeek, Z.ai, Kimi and Mistral prices verified from official sources

    DeepSeek V4 Pro corrected to $0.66 / $1.98 per 1M (official off-peak rate; peak is 2x) and V4 Flash to $0.22 / $0.66. GLM-5.2 corrected to $1.40 / $4.40 (Z.ai). Kimi K3 and Mistral Large 3 / Medium 3.5 confirmed unchanged and now pinned. All official prices are pinned in update-pricing.mjs so the weekly OpenRouter refresh can't overwrite them.

    2026-08-24 — First AI pricing report published

    A new weekly-style report series tracks what changed across the AI pricing catalog and what it means — the first issue covers Gemini 3.7 Flash's launch, DeepSeek V4 Pro going GA, Grok 4.6 matching GPT-5.6 Sol's index at a fraction of the price, and GPT-5.6 Sol's official price being pinned. A generator script (npm run generate:pricing-report) scaffolds future issues from the site's own data.

    2026-08-24 — Prices verified against official provider sources

    API prices re-verified directly against each provider's pricing page. GPT-5.6 Sol corrected to $4 / $20 per 1M (OpenAI; promo through Nov 21 2026) and Gemini 3.7 Flash to $0.75 / $3.75 (Google; promo through Dec 31 2026, then $1.50 / $7.50). Added SuperGrok Lite ($10), SuperGrok Plus ($100) and Google AI Plus ($13.99) plans. Official prices are now pinned so the weekly OpenRouter refresh can't overwrite them.

    2026-08-24 — Rankings hub, subscriptions guide and new editorial

    Added a new /rankings/ hub (front-end design, one-shot coding, cost, speed and overall vibe coding), a /subscriptions/ comparison page, six new head-to-head comparisons, per-model strengths/weaknesses and speed tiers, and 11 new blog posts.

    2026-08-24 — Claude Fable 5 and GLM-5.3 added; catalog re-verified

    Added Anthropic's Claude Fable 5 (intelligence index 62, $10 / $50 per 1M) and Z.ai's GLM-5.3 (max) (index 60, $1.40 / $4.40 per 1M). All API prices re-verified against the OpenRouter catalog and the last-verified dates updated site-wide.

    2026-08-16 — DeepSeek V4 Pro GA (0813) tracked

    DeepSeek V4 Pro moved to its GA 0813 snapshot: $0.435 per 1M input / $0.87 output, with an intelligence index of 53 (up from 45 on the 04-24 preview).

    2026-08-13 — Gemini 3.7 Flash added

    Added Google's newest Flash model at $0.375 / $1.875 per 1M tokens with an intelligence index of 56 — cheaper and higher-indexed than Gemini 3.6 Flash.

    2026-08-12 — Grok 4.6, Kimi K3, GLM-5.2 and Mistral models added

    Added Grok 4.6 (index 61, $2 / $6), Kimi K3 (60, $3 / $15), GLM-5.2 (53, $0.49 / $1.54), Mistral Large 3 and Mistral Medium 3.5.

    2026-08-11 — Full catalog refreshed

    All API prices re-verified against the OpenRouter catalog and the last-verified dates updated site-wide.

    Recently released models

    See every current model's full breakdown on the model pricing page, or compare them all at your own budget in the calculator.

    Advertisement