diff --git a/common/sampling.cpp b/common/sampling.cpp
index 1317024c2c11c..8a4b1f4384f6b 100644
--- a/common/sampling.cpp
+++ b/common/sampling.cpp
@@ -166,7 +166,7 @@ llama_token llama_sampling_sample(
     }
 
     if (ctx_sampling->grammar != NULL) {
-        llama_sample_grammar(ctx_main, &cur_p, ctx_sampling->grammar);
+        llama_sample_grammar(ctx_main, &cur_p, ctx_sampling->grammar, nullptr);
     }
 
     if (temp < 0.0) {
diff --git a/llama.cpp b/llama.cpp
index c5f4053f2fffb..0e87da50a4363 100644
--- a/llama.cpp
+++ b/llama.cpp
@@ -6420,10 +6420,13 @@ struct llama_grammar_candidate {
 // pointer. If an invalid sequence is encountered, returns `llama_partial_utf8.n_remain == -1`.
 static std::pair<std::vector<uint32_t>, llama_partial_utf8> decode_utf8(
         const char         * src,
+        size_t               n_src,
         llama_partial_utf8   partial_start) {
     static const int      lookup[] = { 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 2, 2, 3, 4 };
     const char          * pos      = src;
     std::vector<uint32_t> code_points;
+    // common english strings have the same number of codepoints and bytes. `+ 1` for the terminating 0.
+    code_points.reserve(n_src + 1);
     uint32_t              value    = partial_start.value;
     int                   n_remain = partial_start.n_remain;
 
@@ -6474,6 +6477,13 @@ static std::pair<std::vector<uint32_t>, llama_partial_utf8> decode_utf8(
     return std::make_pair(std::move(code_points), llama_partial_utf8{ value, n_remain });
 }
 
+static std::pair<std::vector<uint32_t>, llama_partial_utf8> decode_utf8(
+        std::string src,
+        llama_partial_utf8 partial_start
+) {
+    return decode_utf8(src.c_str(), src.size(), partial_start);
+}
+
 // returns true iff pos points to the end of one of the definitions of a rule
 static bool llama_grammar_is_end_of_sequence(const llama_grammar_element * pos) {
     switch (pos->type) {
@@ -7096,7 +7106,11 @@ void llama_sample_repetition_penalties(
     }
 }
 
-void llama_sample_grammar(struct llama_context * ctx, llama_token_data_array * candidates, const struct llama_grammar * grammar) {
+void llama_sample_grammar(
+        struct llama_context * ctx,
+        llama_token_data_array * candidates,
+        const struct llama_grammar * grammar,
+        char const * const * pieces) {
     GGML_ASSERT(ctx);
     const int64_t t_start_sample_us = ggml_time_us();
 
@@ -7115,7 +7129,14 @@ void llama_sample_grammar(struct llama_context * ctx, llama_token_data_array * c
 
     for (size_t i = 0; i < candidates->size; ++i) {
         const llama_token id    = candidates->data[i].id;
-        const std::string piece = llama_token_to_piece(ctx, id);
+        std::string piece;
+
+        if (pieces != nullptr && pieces[id] != nullptr) {
+            piece = std::string(pieces[id]);
+        } else {
+            piece = llama_token_to_piece(ctx, id);
+        }
+
         if (id == eos) {
             if (!allow_eos) {
                 candidates->data[i].logit = -INFINITY;
@@ -7123,7 +7144,7 @@ void llama_sample_grammar(struct llama_context * ctx, llama_token_data_array * c
         } else if (piece.empty() || piece[0] == 0) {
             candidates->data[i].logit = -INFINITY;
         } else {
-            candidates_decoded.push_back(decode_utf8(piece.c_str(), grammar->partial_utf8));
+            candidates_decoded.push_back(decode_utf8(piece, grammar->partial_utf8));
             candidates_grammar.push_back({ i, candidates_decoded.back().first.data(), candidates_decoded.back().second });
         }
     }
@@ -7330,7 +7351,7 @@ void llama_grammar_accept_token(struct llama_context * ctx, struct llama_grammar
     const std::string piece = llama_token_to_piece(ctx, token);
 
     // Note terminating 0 in decoded string
-    const auto   decoded     = decode_utf8(piece.c_str(), grammar->partial_utf8);
+    const auto   decoded     = decode_utf8(piece, grammar->partial_utf8);
     const auto & code_points = decoded.first;
     for (auto it = code_points.begin(), end = code_points.end() - 1; it != end; ++it) {
         grammar->stacks = llama_grammar_accept(grammar->rules, grammar->stacks, *it);
diff --git a/llama.h b/llama.h
index 1a62058d1406b..9d6e1599f5988 100644
--- a/llama.h
+++ b/llama.h
@@ -719,10 +719,13 @@ extern "C" {
             "use llama_sample_temp instead");
 
     /// @details Apply constraints from grammar
+    /// @param pieces an array of all null terminated strings obtained from calling llama_token_to_piece for the whole vocab. Can be nullptr in which case they will be computed.
     LLAMA_API void llama_sample_grammar(
             struct llama_context * ctx,
           llama_token_data_array * candidates,
-      const struct llama_grammar * grammar);
+      const struct llama_grammar * grammar,
+              char const * const * pieces);
+
 
     /// @details Mirostat 1.0 algorithm described in the paper https://arxiv.org/abs/2007.14966. Uses tokens instead of words.
     /// @param candidates A vector of `llama_token_data` containing the candidate tokens, their probabilities (p), and log-odds (logit) for the current position in the generated text.