1 more have an end date TBC (not an official deadline; not counted as ending within 7 days). (data)
The largest category is plan usage boosts (7). Every entry links to its official source with a verification time; expired deals move to the archive at the bottom of the page and are never deleted. (data · verified )
Search & filter
⚠ Include region/card unverified
Gemini 4 Argon promotional price (through Nov 30)
Google ✓ Deal exists: officially verified✓ Region & card verifiedNeeds re-check
⚠ Include region/card unverifiedNo end date stated
OpenCode Zen: Beam Free for a limited time (not available in China incl. HK and Macau)
OpenCode ✓ Deal exists: officially verifiedRegion or card unverifiedNeeds re-check
Price Free
Region: restricted (Angola, Belarus, Cambodia, China (incl. Hong Kong and Macau), Cuba, DR Congo, Equatorial Guinea, Iran, Iraq, Laos, Libya, Nigeria, North Korea, Russia, Syria, Turkmenistan, Uzbekistan, Venezuela, Crimea, Donetsk and Luhansk (Ukraine))Real name unverifiedCard unverified
Limited-time free models listed under “Free models” in the official Zen docs (Beam Free has different terms and is listed separately): Big Pickle, GLM-5.3-Flash Free, DeepSeek V4.1 Flash Free, Flint 40B Free, LongCat 2.5 Preview Free, Exo Free, MiMo-V2.6-Flash Free, MiMo-V2.5 Free, Nemotron 3 Ultra Free, Nemotron 3.5 Lightning Free, Nemotron 3.5 Super VL Free and Jev 1.13 Free. The pricing table lists input, output and cached read as Free.
Price
Price Free
Eligibility
Users with an OpenCode account; no region restriction stated (eligibility unverified)
Starts
Varies by model: DeepSeek V4.1 Flash 2026-10-15, GLM 5.3 Flash 2026-10-16; others not stated
Not available in
Unverified
Data use, models, extra sources
Data use
Varies by model (official Privacy section): data from Big Pickle, Exo Free, MiMo-V2.6-Flash Free and MiMo-V2.5 Free may be used to improve the model during the free period; Flint 40B Free likewise, with data retained for 30 days; the three Nemotron free models use NVIDIA free endpoints — trial use only, do not submit personal or confidential data, use is logged for security and to improve NVIDIA products; GLM-5.3-Flash Free, DeepSeek V4.1 Flash Free and LongCat 2.5 Preview Free providers follow zero retention and no training; Jev 1.13 Free has no specific note, so the page-wide statement (zero retention, no training) applies.
Models included (12)
Big Pickle big-pickle · may be used to improve the model
Beam Free is free on OpenCode for a limited time; input, output and cached read are Free; the official X post says 1M context. The provider does not train on your data and retains it for 30 days for safety reasons.
Price
Price Free
Eligibility
OpenCode account holders outside the listed regions (which include China, Hong Kong and Macau) (eligibility unverified)
Starts
2026-10-17 (first official X post)
Not available in
Restricted Angola, Belarus, Cambodia, China (incl. Hong Kong and Macau), Cuba, DR Congo, Equatorial Guinea, Iran, Iraq, Laos, Libya, Nigeria, North Korea, Russia, Syria, Turkmenistan, Uzbekistan, Venezuela, Crimea, Donetsk and Luhansk (Ukraine)
Data use, models, extra sources
Data use
Prompts and inputs are not used for training; data is retained for 30 days for safety reasons.
Models included (1)
Beam Free beam-free · not used for training; kept 30 days
All three models cost $0 for input, output and cache reads. Glyph Cluster is limited to 100 requests per account per day (resets at midnight UTC); starting a session requires at least $1 of credits in the account.
Price
Price Free
Eligibility
All Command Code plans (eligibility unverified)
Starts
Beam 1.0 2026-10-17; others not stated
Not available in
Unverified
Data use, models, extra sources
Data use
For Glyph Cluster and Beam 1.0 the provider may keep prompts and outputs and train on them (official wording); they will not run in a CMD_ZDR=1 session.
Gemini 4 Argon promotional price (through Nov 30) #
Google · Limited-time price cuts✓ Deal exists: officially verifiedNeeds re-check
Paid tier: $2.00 input / $10.00 output per 1M tokens through November 30, 2026; $4.00 / $20.00 starting December 1, 2026. Not available on the free tier.
Gemini 3.8 Flash promotional price (through Dec 31) #
Google · Limited-time price cuts✓ Deal exists: officially verifiedNeeds re-check
Paid tier: $0.75 input / $3.75 output per 1M tokens through December 31, 2026; $1.50 / $7.50 starting January 1, 2027. The free tier lists input and output as free of charge.
MiniMax M3 input $0.60→$0.30, output $2.40→$1.20 and cached read $0.12→$0.06 per 1M tokens (same price at ≤512K and >512K context), so each credit dollar goes 2× further; GOAT allowance $47/month, Pro $57; also applies to top-up credits; applied automatically.
Price
Input · Promo $0.3per 1M tokens
Input · Regular $0.6per 1M tokens
Output · Promo $1.2per 1M tokens
Output · Regular $2.4per 1M tokens
Cached input · Promo $0.06per 1M tokens
Cached input · Regular $0.12per 1M tokens
Eligibility
All Command Code plans and pay-as-you-go credits (eligibility unverified)
Unverified · Paid plans need a payment method (credit/debit card, Alipay or crypto); whether free use needs a card is not stated officially
Region
Unverified · MiniMax M3 is not on the official geo-restriction list (which says restrictions can change); whether the product serves mainland China is not stated officially
Official wording: “GPT-5.6 Sol’s promotional pricing is available at least through November 21, 2026.” Codex credit rates: 100 input / 10 cached input / 500 output credits per 1M tokens. The page does not state the regular rate, so the discount size is not stated.
Price
Unknown
Eligibility
Codex users (ChatGPT plans or credits) (eligibility unverified)
Limited quantity, first come first served, restocked daily at 09:30:00 (UTC+08:00); usage limits: 6,000 requests per 5 hours, 45,000 per week, 90,000 per month; coding tools only, API calls strictly prohibited; no refunds; no sharing
OpenCode Go: $80/month allowance for DeepSeek V4.1 Flash #
OpenCode · Plan usage boosts✓ Deal exists: officially verifiedNeeds re-check
On OpenCode Go ($10/month), DeepSeek V4.1 Flash has an $80 monthly allowance. Peak usage counts at $0.30/$1.20 per 1M input/output tokens, off-peak at $0.15/$0.60; peak is 01:00–04:00 and 06:00–10:00 UTC, Monday–Friday. The model has been upgraded to the 1017 version (official X 2112220734768779441); the allowance is unchanged.
Boosted GLM-5.3 Flash per-model allowance: $10 on Go (the full plan credit), $80 on GOAT, $90 on Pro; per-token prices unchanged.
Price
Unknown
Eligibility
Command Code Go / GOAT / Pro subscribers (eligibility unverified)
Starts
2026-10-17 20:30 UTC+8 (inferred from the end time of the archived three-day $100 boost; official X only says it rises to $80 after three days, no exact time)
Boosted inclusionAI Ling 3.1 Flash per-model allowance: $100 on GOAT, $110 on Pro; per-token prices unchanged. The free version was retired (changelog "retire ling 3.1 flash free"), but the boost remains.
Price
Unknown
Eligibility
Command Code GOAT / Pro subscribers (eligibility unverified)
OpenDesign: doubled DeepSeek V4.1 Flash credits, up to $120/month on Go #
OpenDesign · Plan usage boosts✓ Deal exists: officially verifiedNeeds re-check
DeepSeek V4.1 Flash credits on Go go up to $120/month (from $60); the official X account calls it an ongoing benefit with no scheduled end date. New users can claim free trial credits that expire 72 hours after claiming.
Price
Unknown
Eligibility
OpenDesign Go / Plus / Pro / Max; trial credits for new users (eligibility unverified)
OpenDesign · Plan usage boosts✓ Deal exists: officially verifiedNeeds re-check
Claude Fable 5.1 launch: Pro gets $20 and Max $100 in limited-time bonus credits; GPT-6.1 Sol gets a limited-time usage boost in the Design Plan (2× on Go/Plus, 10× on Pro/Max).
Price
Unknown
Eligibility
Pro / Max (Fable bonus); paid plans (GPT-6.1 Sol boost) (eligibility unverified)
● ActiveEndsNot stated (remaining Cursor contract; up to a year per X post)
Kilo: Kilo Enterprise free for the rest of your Cursor contract #
Kilo · Migration offers✓ Deal exists: officially verifiedNeeds re-check
Official banner: “Cursor loses OpenAI on November 12”; teams paying for Cursor can use Kilo Enterprise free for the rest of their Cursor contract — “up to a year” per the official X post.
Price
Unknown
Eligibility
Teams currently paying for Cursor (eligibility unverified)
The vendor gave an end date without a time zone, and the earliest possible end has passed; not counted as live until confirmed. AoE is never treated as an official extension.
⚠ End date TBCEnd date TBC
The X Subscription (Grok + Cursor + X Premium): 50% off the first month #
xAI · First-month & intro discountsAwaiting official confirmationNeeds re-check
The official announcement lists Plus $8, Premium $30, Super $100 and Ultra $200 per month and says “Switching comes with limited-time offers”. The 50% first-month discount and the 2026-10-27 end date come from a user confirmation and are not stated on an official page.
Each entry must match an official page or the vendor's official account; we show the source link and verification time (UTC+8). Entries confirmed only by users are marked “awaiting official confirmation”. Deals past their end time move to the archive and are never deleted.
Google's pricing page lists $2.00 input / $10.00 output per 1M tokens on the paid tier through November 30, 2026, and $4.00 / $20.00 from December 1, 2026. It is not available on the free tier.
It depends. OpenCode's docs say data from Big Pickle, Flint 40B Free, Exo Free, MiMo-V2.6-Flash Free and MiMo-V2.5 Free may be used to improve the model during the free period, and the three Nemotron free models run on NVIDIA trial endpoints where use is logged; Command Code's docs say the providers of Glyph Cluster and Beam 1.0 may keep and train on prompts and outputs. See each entry's data-use note.
Lite ¥40→¥9.9, Pro ¥200→¥49.9 for up to the first two months, regular price from month three; limited quota, first come first served; one eligibility shared across new purchase, renewal and upgrade; stackable with the 5% referral discount.
When the DeepSeek API is used through DeepSeek Harness, Claude Code or Codex, server-side tool calls hit the Search and Fetch APIs; these are currently free, with per-request billing starting at 00:00 Beijing time on November 1, 2026.
poolside/laguna-s-2.1:free and poolside/laguna-xs-2.1:free are priced at $0. The OpenRouter API lists expiration_date 2026-10-31 on the models themselves (the paid variants expire the same day), not only on the free period.
Price cut “in collab with Xiaomi”, unified across all context lengths; applied automatically. No official end time was given; it ended no later than our observation: we first saw MiMo V2.5 / V2.5 Pro gone from the pricing page DEAL section and price table at 2026-10-23 09:35:40 (UTC+8); still listed at 10-22 09:35.
Price
Unknown
Eligibility
All Command Code plans and pay-as-you-go credits (eligibility unverified)
Official X announced on 2026-10-19 that MiniMax M3.1 Flash was free on OpenCode for 3 days (Zen and Go). On 2026-10-22 the Free variant was removed from the Zen docs and model list; an official repo commit the same day replaced it in Go with paid MiniMax M3.1 Flash ($0.15/$0.60 per 1M tokens). The official commit 074f2fe5 is timed 2026-10-22 22:57:13 (UTC+8); we last confirmed the free model was still available at 2026-10-22 20:39 (UTC+8) and first saw it removed at 23:02 (UTC+8).
Price
Price Free
Eligibility
Users with an OpenCode account; no region restriction stated (eligibility unverified)
Starts
2026-10-19 (official X)
Not available in
Unverified
Ended
expired ·
Data use, models, extra sources
Data use
Official Privacy section: provider follows zero retention and no training.
Models included (1)
MiniMax M3.1 Flash Free minimax-m3.1-flash-free · zero retention
Official X announced on 2026-10-19 that MiniMax M3.1 Flash was free on Command Code for a limited time (all plans, no daily request limit, 1M context). The official docs' deal data gave an expiry of 2026-10-22 10:45 UTC (originally 10-21, extended by one day); after that it was removed from the deal list, and a paid variant is listed at $0.15/$0.60 per 1M tokens. We first observed it gone from the deal list at 2026-10-22 23:38 (UTC+8). The official price table still lists the free row (minimax-m3.1-flash:free, $0); whether it is still usable for free is unverified.
Price
Price Free
Eligibility
All Command Code plans (eligibility unverified)
Starts
2026-10-19 (official X / docs deal data)
Not available in
Unverified
Ended
expired ·
Data use, models, extra sources
Data use
The free variant was zero data retention and ran in CMD_ZDR=1 sessions (official).
Ended Together AI: Muse Spark Open free for a limited time (ended) #
Together AI · Limited-time free models✓ Deal exists: officially verifiedNeeds re-check
The free period ended on 2026-10-18 at 12:00 PM Pacific time; the official page now reads "Promo ended 10/18 12:00PM PST" and the free -promo endpoint is no longer on the Serverless API. The regular model meta-models/Muse-Spark-Open-816B-A42B is billed at $1.25 input / $4.25 output / $0.15 cached per 1M tokens.
Oct 13: the 5-hour and 7-day usage windows were reset for every subscriber (Go: all 12 models; Plus/Pro/Max: 12 Design Plan models). Oct 15: reset again after a DeepSeek outage. OpenDesign says “reset cards you can use when you need them” come next; not launched yet.
Price
Unknown
Eligibility
All subscribers (applied automatically) (eligibility unverified)
Official X: “$10 GOAT → $100 credits” for 3 days, then increased to $80. Current docs show an ongoing boost of $10 Go / $80 GOAT / $90 Pro with no end date.