diff --git a/changes/3987.feature.md b/changes/3987.feature.md
new file mode 100644
index 0000000000..bc6030e201
--- /dev/null
+++ b/changes/3987.feature.md
@@ -0,0 +1 @@
+Two new runtime config options control how the sharding codec coalesces partial-shard reads: `sharding.read.coalesce_max_gap_bytes` (default 1 MiB) and `sharding.read.coalesce_max_bytes` (default 16 MiB). When reading multiple chunks from the same shard, nearby byte ranges are merged into a single request to the store if separated by no more than `coalesce_max_gap_bytes` and the merged read stays within `coalesce_max_bytes`.
diff --git a/docs/user-guide/config.md b/docs/user-guide/config.md
index 8a8fa94c3d..1063b6f35a 100644
--- a/docs/user-guide/config.md
+++ b/docs/user-guide/config.md
@@ -35,6 +35,7 @@ Configuration options include the following:
 - Async and threading options, e.g. `async.concurrency` and `threading.max_workers`
 - Selections of implementations of codecs, codec pipelines and buffers
 - Enabling GPU support with `zarr.config.enable_gpu()`. See GPU support for more.
+- Control request merging when reading multiple chunks from the same shard with `sharding.read.coalesce_max_gap_bytes` and `sharding.read.coalesce_max_bytes`. Reads of nearby chunks are coalesced into a single request to the store when separated by at most `coalesce_max_gap_bytes` and the resulting merged read is no larger than `coalesce_max_bytes`.
 
 For selecting custom implementations of codecs, pipelines, buffers and ndbuffers,
 first register the implementations in the registry and then select them in the config.
diff --git a/src/zarr/codecs/sharding.py b/src/zarr/codecs/sharding.py
index 41f9b7e9c6..37ee69b512 100644
--- a/src/zarr/codecs/sharding.py
+++ b/src/zarr/codecs/sharding.py
@@ -41,6 +41,7 @@
     parse_shapelike,
     product,
 )
+from zarr.core.config import config
 from zarr.core.dtype.npy.int import UInt64
 from zarr.core.indexing import (
     BasicIndexer,
@@ -483,6 +484,8 @@ async def _decode_partial_single(
                 chunk_spec.prototype,
                 chunks_per_shard,
                 all_chunk_coords,
+                max_gap_bytes=config.get("sharding.read.coalesce_max_gap_bytes"),
+                max_coalesced_bytes=config.get("sharding.read.coalesce_max_bytes"),
             )
 
         if shard_dict_maybe is None:
@@ -780,10 +783,16 @@ async def _load_partial_shard_maybe(
         prototype: BufferPrototype,
         chunks_per_shard: tuple[int, ...],
         all_chunk_coords: set[tuple[int, ...]],
+        max_gap_bytes: int,
+        max_coalesced_bytes: int,
     ) -> ShardMapping | None:
         """
         Read chunks from `byte_getter` for the case where the read is less than a full shard.
         Returns a mapping of chunk coordinates to bytes or None.
+
+        `max_gap_bytes` and `max_coalesced_bytes` are forwarded to
+        `Store.get_ranges` to control byte-range coalescing across the requested
+        chunks.
         """
         shard_index = await self._load_shard_index_maybe(byte_getter, chunks_per_shard)
         if shard_index is None:
@@ -807,7 +816,11 @@ async def _load_partial_shard_maybe(
             byte_ranges = [byte_range for _, byte_range in chunk_coord_byte_ranges]
             try:
                 async for group in byte_getter.store.get_ranges(
-                    byte_getter.path, byte_ranges, prototype=prototype
+                    byte_getter.path,
+                    byte_ranges,
+                    prototype=prototype,
+                    max_gap_bytes=max_gap_bytes,
+                    max_coalesced_bytes=max_coalesced_bytes,
                 ):
                     for idx, buf in group:
                         if buf is not None:
diff --git a/src/zarr/core/config.py b/src/zarr/core/config.py
index 7dcbc78e31..3dd9fc18a7 100644
--- a/src/zarr/core/config.py
+++ b/src/zarr/core/config.py
@@ -102,6 +102,12 @@ def enable_gpu(self) -> ConfigSet:
             },
             "async": {"concurrency": 10, "timeout": None},
             "threading": {"max_workers": None},
+            "sharding": {
+                "read": {
+                    "coalesce_max_gap_bytes": 1 << 20,  # 1 MiB
+                    "coalesce_max_bytes": 16 << 20,  # 16 MiB
+                },
+            },
             "json_indent": 2,
             "codec_pipeline": {
                 "path": "zarr.core.codec_pipeline.BatchedCodecPipeline",
diff --git a/tests/test_codecs/test_sharding_unit.py b/tests/test_codecs/test_sharding_unit.py
index 6e022ed9fa..7b07da7749 100644
--- a/tests/test_codecs/test_sharding_unit.py
+++ b/tests/test_codecs/test_sharding_unit.py
@@ -1,3 +1,6 @@
+from typing import TYPE_CHECKING, cast
+from unittest.mock import AsyncMock
+
 import numpy as np
 import pytest
 
@@ -10,9 +13,13 @@
 )
 from zarr.core.buffer import default_buffer_prototype
 from zarr.core.buffer.cpu import Buffer
+from zarr.core.config import config
 from zarr.storage._common import StorePath
 from zarr.storage._memory import MemoryStore
 
+if TYPE_CHECKING:
+    from zarr.core.array import ShardsConfigParam
+
 # ============================================================================
 # _ShardIndex tests
 # ============================================================================
@@ -155,6 +162,8 @@ async def test_load_partial_shard_maybe_index_load_fails() -> None:
         prototype=default_buffer_prototype(),
         chunks_per_shard=(2,),
         all_chunk_coords={(0,)},
+        max_gap_bytes=1 << 20,
+        max_coalesced_bytes=16 << 20,
     )
 
     assert result is None
@@ -187,6 +196,8 @@ async def mock_load_index(
         prototype=default_buffer_prototype(),
         chunks_per_shard=chunks_per_shard,
         all_chunk_coords={(0,), (1,), (2,)},
+        max_gap_bytes=1 << 20,
+        max_coalesced_bytes=16 << 20,
     )
 
     assert result is not None
@@ -220,6 +231,8 @@ async def mock_load_index(
         prototype=default_buffer_prototype(),
         chunks_per_shard=chunks_per_shard,
         all_chunk_coords={(0,), (1,), (2,)},
+        max_gap_bytes=1 << 20,
+        max_coalesced_bytes=16 << 20,
     )
 
     assert result == {}
@@ -251,6 +264,8 @@ async def mock_load_index(
         prototype=default_buffer_prototype(),
         chunks_per_shard=chunks_per_shard,
         all_chunk_coords={(0,), (1,)},
+        max_gap_bytes=1 << 20,
+        max_coalesced_bytes=16 << 20,
     )
 
     assert result is not None
@@ -292,6 +307,8 @@ async def mock_load_index(
         prototype=default_buffer_prototype(),
         chunks_per_shard=chunks_per_shard,
         all_chunk_coords={(0,)},
+        max_gap_bytes=1 << 20,
+        max_coalesced_bytes=16 << 20,
     )
 
     assert result is None
@@ -336,6 +353,8 @@ async def boom(*args: object, **kwargs: object) -> Buffer | None:
             prototype=default_buffer_prototype(),
             chunks_per_shard=chunks_per_shard,
             all_chunk_coords={(0,)},
+            max_gap_bytes=1 << 20,
+            max_coalesced_bytes=16 << 20,
         )
 
 
@@ -368,6 +387,8 @@ async def mock_load_index(
         prototype=default_buffer_prototype(),
         chunks_per_shard=chunks_per_shard,
         all_chunk_coords={(0,), (1,)},
+        max_gap_bytes=1 << 20,
+        max_coalesced_bytes=16 << 20,
     )
 
     assert result is not None
@@ -405,6 +426,8 @@ async def mock_load_index(
         prototype=default_buffer_prototype(),
         chunks_per_shard=chunks_per_shard,
         all_chunk_coords={(0,)},
+        max_gap_bytes=1 << 20,
+        max_coalesced_bytes=16 << 20,
     )
 
     assert result == {}
@@ -486,3 +509,114 @@ def test_is_total_shard_1d() -> None:
     # Partial
     partial_coords: set[tuple[int, ...]] = {(0,), (2,)}
     assert codec._is_total_shard(partial_coords, chunks_per_shard) is False
+
+
+# ============================================================================
+# Coalescing config option tests
+#
+# Asserts that the `sharding.read.coalesce_max_gap_bytes` and
+# `sharding.read.coalesce_max_bytes` config options flow through to
+# `Store.get_ranges` as `max_gap_bytes` / `max_coalesced_bytes` kwargs.
+# ============================================================================
+
+
+def _trigger_partial_shard_read() -> AsyncMock:
+    """Build a sharded array on a mocked `MemoryStore`, trigger a partial-shard
+    read via the public read path, and return the `get_ranges` mock.
+    """
+    import zarr
+    from zarr.codecs.sharding import ShardingCodecIndexLocation
+
+    chunk_shape = (2,)
+    shard_shape = (8,)
+    data = np.arange(8, dtype="int32")
+
+    store = MemoryStore()
+    store_mock = AsyncMock(wraps=store, spec=store.__class__)
+
+    shards: ShardsConfigParam = {
+        "shape": shard_shape,
+        "index_location": ShardingCodecIndexLocation.end,
+    }
+    a = zarr.create_array(
+        StorePath(store_mock),
+        shape=(8,),
+        chunks=chunk_shape,
+        shards=shards,
+        dtype=data.dtype,
+        fill_value=-1,
+    )
+    a[:] = data
+
+    store_mock.reset_mock()
+
+    # Read a strict subset of chunks to take the partial-shard read path.
+    _ = a[0:4]
+
+    return cast(AsyncMock, store_mock.get_ranges)
+
+
+def test_load_partial_shard_forwards_config_to_get_ranges() -> None:
+    """`sharding.read.*` config values are forwarded to `Store.get_ranges`."""
+    with config.set(
+        {
+            "sharding.read.coalesce_max_gap_bytes": 4242,
+            "sharding.read.coalesce_max_bytes": 424242,
+        }
+    ):
+        get_ranges_mock = _trigger_partial_shard_read()
+
+    assert get_ranges_mock.call_count >= 1
+    for call in get_ranges_mock.call_args_list:
+        kwargs = call.kwargs
+        assert kwargs["max_gap_bytes"] == 4242
+        assert kwargs["max_coalesced_bytes"] == 424242
+
+
+def test_load_partial_shard_uses_config_defaults() -> None:
+    """Without explicit config, defaults from `zarr.config` are forwarded."""
+    get_ranges_mock = _trigger_partial_shard_read()
+
+    assert get_ranges_mock.call_count >= 1
+    for call in get_ranges_mock.call_args_list:
+        kwargs = call.kwargs
+        assert kwargs["max_gap_bytes"] == config.get("sharding.read.coalesce_max_gap_bytes")
+        assert kwargs["max_coalesced_bytes"] == config.get("sharding.read.coalesce_max_bytes")
+
+
+async def test_load_partial_shard_explicit_kwargs_passthrough(
+    monkeypatch: pytest.MonkeyPatch,
+) -> None:
+    """`_load_partial_shard_maybe` forwards its explicit kwargs to `get_ranges`."""
+    codec = ShardingCodec(chunk_shape=(2,))
+    chunks_per_shard = (4,)
+
+    index = _ShardIndex.create_empty(chunks_per_shard)
+    index.set_chunk_slice((0,), slice(0, 100))
+    index.set_chunk_slice((2,), slice(200, 300))
+
+    store = MemoryStore()
+    await store.set("shard", Buffer.from_bytes(b"x" * 300))
+    store_mock = AsyncMock(wraps=store, spec=store.__class__)
+    byte_getter = StorePath(store_mock, "shard")
+
+    async def mock_load_index(
+        self: ShardingCodec, byte_getter: StorePath, cps: tuple[int, ...]
+    ) -> _ShardIndex:
+        return index
+
+    monkeypatch.setattr(ShardingCodec, "_load_shard_index_maybe", mock_load_index)
+
+    await codec._load_partial_shard_maybe(
+        byte_getter=byte_getter,
+        prototype=default_buffer_prototype(),
+        chunks_per_shard=chunks_per_shard,
+        all_chunk_coords={(0,), (2,)},
+        max_gap_bytes=12345,
+        max_coalesced_bytes=67890,
+    )
+
+    store_mock.get_ranges.assert_called_once()
+    kwargs = store_mock.get_ranges.call_args.kwargs
+    assert kwargs["max_gap_bytes"] == 12345
+    assert kwargs["max_coalesced_bytes"] == 67890
diff --git a/tests/test_config.py b/tests/test_config.py
index 4e293e968f..622305880e 100644
--- a/tests/test_config.py
+++ b/tests/test_config.py
@@ -59,6 +59,12 @@ def test_config_defaults_set() -> None:
                 },
                 "async": {"concurrency": 10, "timeout": None},
                 "threading": {"max_workers": None},
+                "sharding": {
+                    "read": {
+                        "coalesce_max_gap_bytes": 1 << 20,
+                        "coalesce_max_bytes": 16 << 20,
+                    },
+                },
                 "json_indent": 2,
                 "codec_pipeline": {
                     "path": "zarr.core.codec_pipeline.BatchedCodecPipeline",