What IR64 actually is

IR64 was bred at the International Rice Research Institute and first released in 1985 by the Philippine Seed Board. The cross behind it, IR5657-33-2-1 / IR2061-465-1-5-5, was made in early 1977 and designated IR18348; the selected line IR18348-36-3-3 became IR64. Its pedigree traces back to nineteen traditional varieties, and the traditional variety BPI-121-407 is considered the most likely source of its grain quality.

It became what breeders call a mega variety. It was released in twelve countries — the Philippines first, then Bhutan, Burkina Faso, Cambodia, China, Ecuador, the Gambia, India, Indonesia, Mauritania, Mozambique and Vietnam — and in several of those under a local name rather than its IRRI designation:

It reached 8 million hectares by 1995 and over 10 million hectares by around 2000, running at an estimated 9 to 10 million hectares annually at its height. What made it travel was the combination: it was the first variety to bring together intermediate amylose content, soft gel consistency, intermediate gelatinisation temperature, translucent grains and long slender grains, and it carried early maturity, high yield and resistance to blast, bacterial blight and brown planthopper. It also mills unusually well, with head rice recovery among the highest recorded. In several countries it became the quality standard against which other varieties were judged.

The variety-versus-specification problem

Here is the part that affects what arrives in your warehouse, and it is a disclosure against our own commercial interest.

The peer-reviewed review of IR64 by Mackill and Khush, published in the journal Rice in 2018, states plainly that the variety’s area has declined over the past decade and that it has been succeeded by a new generation of high-quality varieties that are mostly its own progeny and relatives. The detail by country is specific:

• Philippines — area declined from 2000 to 2007, partly because of tungro disease, and it was replaced by newer varieties in the early 2000s.
• Indonesia — it had covered over 40 percent of the total rice area for roughly a decade before being replaced by Ciherang.
• India — it remains popular in the north, but the area is gradually declining. Breeder seed production was above 3 percent in 2015, implying roughly 2 to 3 million hectares annually across 1998 to 2006.

Set that against the way the name is used in trade, where “IR64” functions as shorthand for long-grain indica of a particular length, translucency, milling behaviour and cooking profile. Both things are true at once. The consequence is simple and worth stating without decoration:

One source note, consistent with how we handle this series. The CGIAR innovation page on IR64, published in 2021, describes the variety in the present tense as cultivated in most rice-growing countries and makes no mention of declining area. The 2018 peer-reviewed review says the area has declined and names the replacements. We are relying on the review rather than the innovation page, because it is peer-reviewed, country-specific and gives its evidence — but you should know the two sources read differently rather than discover it later.

Amylose: the number that predicts the pot

If you only take one technical figure from this page, take this one. Amylose content is described in the literature as the major determinant of rice texture, and it predicts cooking behaviour better than any other single measurement.

FAO classifies milled rice by amylose content as follows:

IR64 is an intermediate-amylose rice, measured at 21.2 percent apparent amylose in a peer-reviewed survey published in PLOS ONE, alongside intermediate gelatinisation temperature and soft gel consistency. That combination is the whole commercial point of the variety, and it explains the behaviour buyers actually notice:

• Grains stay separate and hold their shape through service, which is why it performs in institutional catering, canteens and programme feeding.
• It tolerates holding and reheating better than a low-amylose rice, which clings and breaks down.
• It parboils well. FAO notes that higher-amylose rices with medium to soft gel consistency are preferred across most of South Asia precisely because they suit parboiling — and parboiled IR64 is a very large share of what India ships to West Africa and Bangladesh.
• It will disappoint anyone wanting a soft, clingy grain. If your market expects rice that mounds on the plate, intermediate amylose is the wrong material and no amount of milling specification will fix it.

Two honest caveats. Amylose varies with growing environment, season and analytical method, so treat 21.2 percent as representative of the class rather than a constant you can hold a supplier to without testing. And because trade IR64 is partly a specification rather than a pedigree, as above, the amylose of a given lot is a property of that lot. If your process depends on it — a parboiling line, a ready-meal, an extruded product — ask for a measured value on the shipped lot rather than a varietal average.

The forms IR64 is sold in

Codex Standard 198 defines the broken fractions precisely, which is worth knowing because “broken” is argued over more than any other line in a rice contract:

Grain size and tolerances: contract on Codex, not on adjectives

Because the variety name carries less information than buyers assume, the grain description has to carry more. Codex Standard 198 is the right framework: it is international, both parties can look it up, and it removes the argument about what a word means.

Length classification

Codex offers three alternative options, and a contract should say which one it is using:

IR64 is described in the literature as a typical long-grain variety with long slender translucent grains, but no authoritative source gives a canonical millimetre figure for it — and given how the name is used in trade, a canonical figure would be misleading anyway. Ask for the measured average milled length and the length-to-width ratio on the lot being offered, and write both into the contract. That single step replaces the entire argument about whether something “is really IR64”.

Codex maximum tolerances

These are the international ceilings. Most commercial contracts tighten them substantially, and you should treat them as the outer limit of what can lawfully be called the product rather than as a target:

Note the parboiled columns. Parboiling raises the allowance for heat-damaged kernels from 3.0 to 6.0 percent in milled rice and introduces pecks as a defect category, because the process itself can produce both. If you are buying parboiled IR64 and applying a raw-white defect schedule to it, you have written a contract that the material cannot meet.

Where Indian IR64 goes, and two markets that just closed

This is where the widely republished “top rice importing countries” lists will cost you money if you use them to plan.

India’s non-basmati destinations, 2025-26

APEDA reports 15.01 million MT worth Rs 51,892.32 crore, about US$5.82 billion:

The two largest importers both switched themselves off

On 2024 data, Indonesia and the Philippines were the world’s number one and number two rice importers. Both then contracted sharply, and most published lists have not caught up:

• Indonesia imported no rice in 2025. In December 2025 Agriculture Minister Amran Sulaiman, who also heads the National Food Agency, announced a Bulog stockpile of 3.7 million tonnes, the largest since Bulog was established in 1967, and stated that Indonesia had stopped importing rice. In January 2026 the National Food Agency confirmed no rice or sugar imports for consumption, and no feed corn imports, in 2026, citing national stocks of about 12.5 million tonnes — close to five months of demand — against forecast production of 34.7 million tonnes. Indonesia’s 2025 rice import value fell by roughly 92 percent.
• The Philippines suspended and then reopened. A 60-day suspension of regular milled and well-milled rice imports took effect on 1 September 2025 and was extended to the end of the year. Imports fell about 30 percent for 2025, from 4.8 million MT in 2024 to 3.37 million MT. The market reopened in January 2026 with Agriculture Secretary Francisco Tiu Laurel Jr. indicating 3.6 to 3.8 million MT planned for the year, and the import tariff raised from 15 to 20 percent. Sources differ on whether the new rate took effect on 1 January or 16 January 2026, so verify the rate applicable at your shipment date rather than relying on either figure.

Meanwhile the West African markets that actually buy Indian IR64 grew through the same period. On 2025 figures, Côte d’Ivoire rose 10.2 percent to about US$1.11 billion, Benin rose 21.8 percent to about US$864 million, Senegal rose 11.9 percent to about US$580 million, and Guinea-Bissau rose 27.4 percent.

Arsenic: the parboiled limit is higher, and that is not leniency

Because parboiled IR64 is such a large share of Indian shipments, this one matters here more than for rice generally. Under Regulation (EU) 2023/465 amending Regulation (EU) 2023/915, the maximum levels for inorganic arsenic are:

Parboiled and husked rice get 0.25 mg/kg where white rice gets 0.15. That reads like a concession and it is not one. Inorganic arsenic concentrates in the bran layer. Milling strips the bran and takes much of the arsenic with it, which is why polished white carries the tighter number — the material is genuinely cleaner. Husked rice keeps the bran, and parboiling drives some of the arsenic inward from the outer layers into the endosperm, where milling can no longer remove it. The limit is higher because the material carries more, not because the risk is lower. Switching from raw white IR64 to parboiled IR64 buys you no headroom.

Infant food overrides form entirely: at 0.10 mg/kg, rice destined for infant and young-child food faces a tighter limit than white rice itself, and parboiled rice intended for later conversion into baby food is the hardest combination in the table. Say so at enquiry, not after shipment, because it changes which lots can be offered. Cadmium is separately limited at 0.15 mg/kg on a row that includes rice.

EU border controls: rice is listed, under the lighter regime

Rice from India is subject to increased official controls, and the distinction between the two annexes is worth money.

Rice is listed under CN 1006 in Annex I of Regulation (EU) 2019/1793 for pesticide residues at a 10 percent control rate. We verified this against the India entries as amended by Implementing Regulation (EU) 2026/1206, in force from 30 June 2026. Rice remains the only cereal listed for India in either annex.

So rice is checked more often than average, and that is all. There is no Annex IV certificate to obtain, no four-month clock to run, no code to match across three documents. If a freight forwarder or agent tells you Indian rice needs the full Annex II pack, they are applying the spice rules to a cereal. Two standard caveats: the 10 percent rate is a sampling frequency rather than a guarantee of passage, and the annexes are reviewed roughly twice a year, so check the position at your shipment date.

Policy risk, which neither of us controls

Everything is open at the moment. The path here matters more than the current state: India changed its rice export rules five times in roughly thirty months — an additional duty on non-basmati white rice and a broken-rice export ban in September 2022; a full white-rice ban with a 20 percent parboiled duty in September 2023; the ban replaced by a US$490/MT minimum export price with the parboiled duty cut to 10 percent in September 2024; both removed in October 2024; and the broken-rice ban lifted on 7 March 2025, when ITC HS 10064000 moved from Prohibited to Free with immediate effect.

The honest sentence is not that India is open for business. It is that Indian export policy is the single largest risk in buying Indian rice, and neither we nor you control it. Price volatility in, keep a second origin qualified, and write force majeure and act-of-government clauses that address export prohibition explicitly rather than leaving it to general wording. We would rather negotiate that with you now than find the gap on the day a notification issues.

Packing: private labelling from 1 kg to 50 kg, and bulk shipments

RNG Agro Exports provides private labelling from 1 kg pack size to 50 kg pack size, and we do bulk shipments as well. This is a real part of the business rather than a line in a brochure, so here is how the formats map to markets — because pack size changes the specification, not just the appearance.

Private label work covers artwork and printing, language versions, nutrition and ingredient panels, barcodes, and country-specific labelling requirements, which differ sharply between, say, a European retail pack and a West African one. Tell us the destination market and the channel when you enquire. A 1 kg retail bag and a 50 kg woven sack are not the same product in two sizes: moisture tolerance, barrier, pallet configuration and shelf-life declaration all differ, and so does the price.

Specification: what should be in the contract

If you want it organic

Certified organic IR64 adds three documentation layers on top of the specification above:

The usual warning applies: organic is not an analytical zero. Organic certification governs what may be applied, not what may be detected. And one point deserves saying outright here: organic certification does nothing for arsenic. Inorganic arsenic in rice comes from soil and irrigation water taken up by the plant, not from an applied input, so the organic certificate is silent on the contaminant most likely to stop a rice consignment.

Frequently asked questions

Is the IR64 I am offered genuinely the IR64 variety?

Often it is a grain specification rather than a verified pedigree, and you should assume so unless you have contracted otherwise. IR64 was bred at IRRI and released in 1985 by the Philippine Seed Board, from the cross IR5657-33-2-1/IR2061-465-1-5-5 made in early 1977. It went on to be released in twelve countries, under local names in several of them, including FKR42 in Burkina Faso, NIAP11 in Ecuador and OM89 in Vietnam. It reached 8 million hectares by 1995 and over 10 million hectares by around 2000. But the peer-reviewed review by Mackill and Khush in the journal Rice in 2018 states that its area has declined over the past decade and that it has been succeeded by a new generation of varieties that are mostly its own progeny and relatives. In India it remains popular in the north while its area gradually declines. So a large share of what moves internationally as IR64 is long-grain indica matching IR64’s length, translucency and cooking profile rather than certified seed stock. For most staple buyers that is not a problem, because the specification is what performs in the pot. It only becomes a problem if you assumed otherwise. If you need the actual variety, say so in the contract and ask how it will be evidenced.

What does IR64’s amylose content mean for how it cooks?

It is the single most useful number for predicting behaviour in the pot, and IR64 sits in the intermediate band. Amylose is described in the literature as the major determinant of rice texture. FAO classifies milled rice as waxy at 0 to 5 percent, very low at 5.1 to 12.0, low at 12.1 to 20.0, intermediate at 20.1 to 25.0 and high above 25.0. IR64 is classified as intermediate, and was measured at 21.2 percent apparent amylose in a peer-reviewed survey published in PLOS ONE. Intermediate and high amylose rices cook firm and fluffy with grains that stay separate; low amylose and waxy rices cook soft, moist and sticky. That is why IR64 holds up in institutional catering, tolerates holding and reheating, and takes parboiling well, and it is also why it will disappoint anyone expecting a soft, clingy grain. Note that amylose varies with growing environment and season, so treat 21.2 percent as a representative figure for the class rather than a constant, and ask for a measured value if your process depends on it.

Which markets actually buy Indian IR64 rice?

Overwhelmingly West Africa and Bangladesh, and the widely republished lists of the world’s largest rice importers will mislead you on this. APEDA reports India exported 15.01 million MT of non-basmati rice worth Rs 51,892.32 crore, about US$5.82 billion, in 2025-26, with top destinations by share being Benin 11.5 percent, Bangladesh 9.3 percent, Guinea 6.6 percent, Togo 6.6 percent and Côte d’Ivoire 5.4 percent. Four of the five are West African. Meanwhile the two markets that topped the 2024 world import tables have both contracted sharply: Indonesia imported no rice at all in 2025 and confirmed no consumption imports for 2026, and the Philippines suspended imports from 1 September 2025 to the end of the year and fell about 30 percent for the year, from 4.8 million MT in 2024 to 3.37 million MT, before reopening in January 2026 with 3.6 to 3.8 million MT planned. Note separately that Saudi Arabia’s large headline number is mostly basmati: it is India’s top basmati destination at 16.6 percent of basmati exports and does not appear in the non-basmati top five at all.

What should I specify in an IR64 contract?

Specify the grain rather than relying on the name, and use Codex Standard 198 as the framework because it is the one reference both sides can check. State the form, meaning parboiled, raw white, husked or broken, and the broken percentage. State average milled grain length in millimetres and the length to width ratio, since Codex classifies milled rice as long grain at 6.6 mm or more by length, or at a ratio of 3.0 or more. State moisture, where the Codex maximum is 15 percent by mass but lower figures are normal for long transit. State maximum chalky, damaged, heat-damaged, immature, red and red-streaked kernels; Codex allows up to 11.0 percent chalky and 3.0 percent damaged in milled rice, and 6.0 percent heat-damaged in parboiled milled rice, and most commercial contracts tighten these considerably. State extraneous matter, with Codex at 0.5 percent organic and 0.1 percent inorganic for milled rice. Add crop year, degree of milling against an agreed reference sample, inorganic arsenic against the correct regulatory category, a pesticide residue screen matched to the destination, packing and labelling, fumigation method, and a clause covering export prohibition or duty imposed after the contract date.

What pack sizes and private label options do you offer?

RNG Agro Exports provides private labelling from 1 kg pack size to 50 kg pack size, and we do bulk shipments as well. In practice that spans 1 kg to 5 kg consumer packs for supermarket and grocery retail, diaspora stores and e-commerce; 10 kg to 25 kg packs for foodservice and for West African and South Asian household retail where larger household packs are the norm; 50 kg bags for wholesale, institutional supply and relief and programme tenders; and bulk or containerised shipments for traders, millers and repackers applying their own brand at destination. Private label work covers artwork, language versions, nutrition and ingredient panels, barcodes and country-specific labelling requirements. Tell us the destination market and the channel when you enquire, because the pack format changes the specification: moisture tolerance, barrier film, pallet configuration and shelf-life declaration are not the same for a 1 kg retail bag as for a 50 kg woven sack.

• Often it is a grain specification rather than a verified pedigree, and you should assume so unless you have contracted otherwise. IR64 was bred at IRRI and released in 1985 by the Philippine Seed Board, from the cross IR5657-33-2-1/IR2061-465-1-5-5 made in early 1977. It went on to be released in twelve countries, under local names in several of them, including FKR42 in Burkina Faso, NIAP11 in Ecuador and OM89 in Vietnam. It reached 8 million hectares by 1995 and over 10 million hectares by around 2000. But the peer-reviewed review by Mackill and Khush in the journal Rice in 2018 states that its area has declined over the past decade and that it has been succeeded by a new generation of varieties that are mostly its own progeny and relatives. In India it remains popular in the north while its area gradually declines. So a large share of what moves internationally as IR64 is long-grain indica matching IR64’s length, translucency and cooking profile rather than certified seed stock. For most staple buyers that is not a problem, because the specification is what performs in the pot. It only becomes a problem if you assumed otherwise. If you need the actual variety, say so in the contract and ask how it will be evidenced.
• It is the single most useful number for predicting behaviour in the pot, and IR64 sits in the intermediate band. Amylose is described in the literature as the major determinant of rice texture. FAO classifies milled rice as waxy at 0 to 5 percent, very low at 5.1 to 12.0, low at 12.1 to 20.0, intermediate at 20.1 to 25.0 and high above 25.0. IR64 is classified as intermediate, and was measured at 21.2 percent apparent amylose in a peer-reviewed survey published in PLOS ONE. Intermediate and high amylose rices cook firm and fluffy with grains that stay separate; low amylose and waxy rices cook soft, moist and sticky. That is why IR64 holds up in institutional catering, tolerates holding and reheating, and takes parboiling well, and it is also why it will disappoint anyone expecting a soft, clingy grain. Note that amylose varies with growing environment and season, so treat 21.2 percent as a representative figure for the class rather than a constant, and ask for a measured value if your process depends on it.
• Overwhelmingly West Africa and Bangladesh, and the widely republished lists of the world’s largest rice importers will mislead you on this. APEDA reports India exported 15.01 million MT of non-basmati rice worth Rs 51,892.32 crore, about US$5.82 billion, in 2025-26, with top destinations by share being Benin 11.5 percent, Bangladesh 9.3 percent, Guinea 6.6 percent, Togo 6.6 percent and Côte d’Ivoire 5.4 percent. Four of the five are West African. Meanwhile the two markets that topped the 2024 world import tables have both contracted sharply: Indonesia imported no rice at all in 2025 and confirmed no consumption imports for 2026, and the Philippines suspended imports from 1 September 2025 to the end of the year and fell about 30 percent for the year, from 4.8 million MT in 2024 to 3.37 million MT, before reopening in January 2026 with 3.6 to 3.8 million MT planned. Note separately that Saudi Arabia’s large headline number is mostly basmati: it is India’s top basmati destination at 16.6 percent of basmati exports and does not appear in the non-basmati top five at all.
• Specify the grain rather than relying on the name, and use Codex Standard 198 as the framework because it is the one reference both sides can check. State the form, meaning parboiled, raw white, husked or broken, and the broken percentage. State average milled grain length in millimetres and the length to width ratio, since Codex classifies milled rice as long grain at 6.6 mm or more by length, or at a ratio of 3.0 or more. State moisture, where the Codex maximum is 15 percent by mass but lower figures are normal for long transit. State maximum chalky, damaged, heat-damaged, immature, red and red-streaked kernels; Codex allows up to 11.0 percent chalky and 3.0 percent damaged in milled rice, and 6.0 percent heat-damaged in parboiled milled rice, and most commercial contracts tighten these considerably. State extraneous matter, with Codex at 0.5 percent organic and 0.1 percent inorganic for milled rice. Add crop year, degree of milling against an agreed reference sample, inorganic arsenic against the correct regulatory category, a pesticide residue screen matched to the destination, packing and labelling, fumigation method, and a clause covering export prohibition or duty imposed after the contract date.

A significant share of the rice moving internationally as IR64 is a grain specification rather than certified seed stock. It may be IR64, it may be one of its descendants, it may be a comparable long-grain indica selected to match. For the overwhelming majority of staple buyers this is not a defect, because what you are buying is a performance profile and the profile is what the specification describes. It becomes a problem only in two situations: when you assumed you were buying verified varietal identity and priced it accordingly, and when your end use genuinely depends on the variety — seed multiplication, a research programme, a product claim on pack, or a contract that names the variety as a warranty. If you are in one of those situations, say so, write it into the contract, and ask us how it will be evidenced. If you are not, stop paying attention to the name and start specifying the grain. The next three sections are about how to do that.