diff --git a/internal/gateway/chat.go b/internal/gateway/chat.go
index 8407da4..71f43a6 100644
--- a/internal/gateway/chat.go
+++ b/internal/gateway/chat.go
@@ -575,8 +575,14 @@ func recordChatUsage(rec *Req, req *types.ChatRequest, resp *types.UnifiedRespon
 	// read it back so the request record carries the hit/miss split.
 	if d := resp.TokenUsage.PromptTokensDetails; d != nil && d.CachedTokens > 0 {
 		rec.CacheHit = int64(d.CachedTokens)
+	}
+	if resp.TokenUsage.PromptTokensDetails != nil {
+		// Upstream reported cache details (even a 0 hit) — tag the row so
+		// the UI can show 0% rather than “—”.
+		rec.CacheReported = true
 	} else if resp.TokenUsage.PromptCacheHit > 0 {
 		rec.CacheHit = int64(resp.TokenUsage.PromptCacheHit)
+		rec.CacheReported = true
 	}
 	if resp.TokenUsage.PromptCacheMiss > 0 {
 		rec.CacheMiss = int64(resp.TokenUsage.PromptCacheMiss)
@@ -761,10 +767,14 @@ func (g *Gateway) pumpStream(w http.ResponseWriter, rec *Req, chunks <-chan type
 		tut = lastUsage
 		// Write the upstream's cache accounting back onto the request record
 		// so the audit trail carries the hit/miss split for streaming too.
-		if d := tut.PromptTokensDetails; d != nil && d.CachedTokens > 0 {
-			rec.CacheHit = int64(d.CachedTokens)
+		if d := tut.PromptTokensDetails; d != nil {
+			rec.CacheReported = true
+			if d.CachedTokens > 0 {
+				rec.CacheHit = int64(d.CachedTokens)
+			}
 		} else if tut.PromptCacheHit > 0 {
 			rec.CacheHit = int64(tut.PromptCacheHit)
+			rec.CacheReported = true
 		}
 		if tut.PromptCacheMiss > 0 {
 			rec.CacheMiss = int64(tut.PromptCacheMiss)
diff --git a/internal/gateway/stats.go b/internal/gateway/stats.go
index 2cef29e..2f2505f 100644
--- a/internal/gateway/stats.go
+++ b/internal/gateway/stats.go
@@ -35,6 +35,10 @@ type Req struct {
 	// Both 0 = upstream gave no cache data.
 	CacheHit  int64 `json:"cache_hit_tokens,omitempty"`
 	CacheMiss int64 `json:"cache_miss_tokens,omitempty"`
+	// CacheReported marks that the upstream usage reported cache
+	// accounting at all (even when the hit count is 0). The WebUI shows
+	// "0%" instead of "—" for such rows.
+	CacheReported bool `json:"cache_reported,omitempty"`
 	OK    bool  `json:"ok"`
 	// Status http status code
 	Status int `json:"status"`
diff --git a/internal/gateway/ui/index.html b/internal/gateway/ui/index.html
index 52c1b87..c34898d 100644
--- a/internal/gateway/ui/index.html
+++ b/internal/gateway/ui/index.html
@@ -1990,10 +1990,11 @@
       // the upstream reported cache accounting, otherwise an em dash.
       function cacheCell(r) {
         const hit = r.cache_hit_tokens || 0;
-        if (!hit && !(r.cache_miss_tokens > 0)) return '<span class="muted">—</span>';
+        if (!hit && !(r.cache_miss_tokens > 0) && !r.cache_reported)
+          return '<span class="muted" title="上游未报告缓存数据">—</span>';
         const prompt = r.prompt_tokens || 0;
         const pct = prompt > 0 ? Math.round((hit * 100) / prompt) : 0;
-        return `<span class="tag ${pct >= 50 ? "tag-green" : pct > 0 ? "tag-amber" : ""}" title="命中 ${hit} / 未命中 ${r.cache_miss_tokens || 0}">${pct}%</span>`;
+        return `<span class="tag ${pct >= 50 ? "tag-green" : "tag-amber"}" title="命中 ${hit} / 未命中 ${r.cache_miss_tokens || 0}">${pct}%</span>`;
       }
       function showModelConfig(srcName, model) {
         const el = $("#conncfg");
diff --git a/internal/lua/adapters/agentrouter.lua b/internal/lua/adapters/agentrouter.lua
index 4decf4b..48cdfe4 100644
--- a/internal/lua/adapters/agentrouter.lua
+++ b/internal/lua/adapters/agentrouter.lua
@@ -53,7 +53,7 @@ function adapter.transform_response(raw_body)
         unified.token_usage.completion = resp.usage.completion_tokens or 0
         unified.token_usage.total = resp.usage.total_tokens or 0
         local hit = 0
-        if type(resp.usage.prompt_tokens_details) == "table" and (resp.usage.prompt_tokens_details.cached_tokens or 0) > 0 then
+        if type(resp.usage.prompt_tokens_details) == "table" and resp.usage.prompt_tokens_details.cached_tokens ~= nil then
             hit = resp.usage.prompt_tokens_details.cached_tokens
             unified.token_usage.prompt_tokens_details = { cached_tokens = hit }
         end
@@ -107,7 +107,7 @@ function adapter.transform_stream_chunk(raw_chunk)
             completion = chunk.usage.completion_tokens or chunk.usage.completion or 0,
             total = chunk.usage.total_tokens or chunk.usage.total or 0,
         }
-        if type(chunk.usage.prompt_tokens_details) == "table" and (chunk.usage.prompt_tokens_details.cached_tokens or 0) > 0 then
+        if type(chunk.usage.prompt_tokens_details) == "table" and chunk.usage.prompt_tokens_details.cached_tokens ~= nil then
             uses.prompt_tokens_details = { cached_tokens = chunk.usage.prompt_tokens_details.cached_tokens }
         elseif (chunk.usage.prompt_cache_hit_tokens or 0) > 0 then
             uses.prompt_cache_hit_tokens = chunk.usage.prompt_cache_hit_tokens
diff --git a/internal/lua/adapters/deepseek.lua b/internal/lua/adapters/deepseek.lua
index 9f65e71..310c59e 100644
--- a/internal/lua/adapters/deepseek.lua
+++ b/internal/lua/adapters/deepseek.lua
@@ -51,7 +51,7 @@ function adapter.transform_response(raw_body)
         unified.token_usage.total = resp.usage.total_tokens or 0
         unified.token_usage.prompt_cache_hit_tokens = resp.usage.prompt_cache_hit_tokens or 0
         unified.token_usage.prompt_cache_miss_tokens = resp.usage.prompt_cache_miss_tokens or 0
-        if resp.usage.prompt_tokens_details and resp.usage.prompt_tokens_details.cached_tokens then
+        if resp.usage.prompt_tokens_details and resp.usage.prompt_tokens_details.cached_tokens ~= nil then
             unified.token_usage.prompt_tokens_details = { cached_tokens = resp.usage.prompt_tokens_details.cached_tokens }
         end
     end
@@ -100,7 +100,7 @@ function adapter.transform_stream_chunk(raw_chunk)
             prompt_cache_hit_tokens = chunk.usage.prompt_cache_hit_tokens or 0,
             prompt_cache_miss_tokens = chunk.usage.prompt_cache_miss_tokens or 0,
         }
-        if chunk.usage.prompt_tokens_details and chunk.usage.prompt_tokens_details.cached_tokens then
+        if chunk.usage.prompt_tokens_details and chunk.usage.prompt_tokens_details.cached_tokens ~= nil then
             uses.prompt_tokens_details = { cached_tokens = chunk.usage.prompt_tokens_details.cached_tokens }
         end
     end
diff --git a/internal/lua/adapters/github.lua b/internal/lua/adapters/github.lua
index a7ace2c..0db6d9b 100644
--- a/internal/lua/adapters/github.lua
+++ b/internal/lua/adapters/github.lua
@@ -34,7 +34,7 @@ function adapter.transform_response(raw_body)
         unified.token_usage.completion = resp.usage.completion_tokens or 0
         unified.token_usage.total = resp.usage.total_tokens or 0
         local hit = 0
-        if type(resp.usage.prompt_tokens_details) == "table" and (resp.usage.prompt_tokens_details.cached_tokens or 0) > 0 then
+        if type(resp.usage.prompt_tokens_details) == "table" and resp.usage.prompt_tokens_details.cached_tokens ~= nil then
             hit = resp.usage.prompt_tokens_details.cached_tokens
             unified.token_usage.prompt_tokens_details = { cached_tokens = hit }
         end
@@ -86,7 +86,7 @@ function adapter.transform_stream_chunk(raw_chunk)
             completion = chunk.usage.completion_tokens or chunk.usage.completion or 0,
             total = chunk.usage.total_tokens or chunk.usage.total or 0,
         }
-        if type(chunk.usage.prompt_tokens_details) == "table" and (chunk.usage.prompt_tokens_details.cached_tokens or 0) > 0 then
+        if type(chunk.usage.prompt_tokens_details) == "table" and chunk.usage.prompt_tokens_details.cached_tokens ~= nil then
             uses.prompt_tokens_details = { cached_tokens = chunk.usage.prompt_tokens_details.cached_tokens }
         elseif (chunk.usage.prompt_cache_hit_tokens or 0) > 0 then
             uses.prompt_cache_hit_tokens = chunk.usage.prompt_cache_hit_tokens
diff --git a/internal/lua/adapters/groq.lua b/internal/lua/adapters/groq.lua
index 88ea0ad..d8858bc 100644
--- a/internal/lua/adapters/groq.lua
+++ b/internal/lua/adapters/groq.lua
@@ -33,7 +33,7 @@ function adapter.transform_response(raw_body)
         unified.token_usage.completion = resp.usage.completion_tokens or 0
         unified.token_usage.total = resp.usage.total_tokens or 0
         local hit = 0
-        if type(resp.usage.prompt_tokens_details) == "table" and (resp.usage.prompt_tokens_details.cached_tokens or 0) > 0 then
+        if type(resp.usage.prompt_tokens_details) == "table" and resp.usage.prompt_tokens_details.cached_tokens ~= nil then
             hit = resp.usage.prompt_tokens_details.cached_tokens
             unified.token_usage.prompt_tokens_details = { cached_tokens = hit }
         end
@@ -85,7 +85,7 @@ function adapter.transform_stream_chunk(raw_chunk)
             completion = chunk.usage.completion_tokens or chunk.usage.completion or 0,
             total = chunk.usage.total_tokens or chunk.usage.total or 0,
         }
-        if type(chunk.usage.prompt_tokens_details) == "table" and (chunk.usage.prompt_tokens_details.cached_tokens or 0) > 0 then
+        if type(chunk.usage.prompt_tokens_details) == "table" and chunk.usage.prompt_tokens_details.cached_tokens ~= nil then
             uses.prompt_tokens_details = { cached_tokens = chunk.usage.prompt_tokens_details.cached_tokens }
         elseif (chunk.usage.prompt_cache_hit_tokens or 0) > 0 then
             uses.prompt_cache_hit_tokens = chunk.usage.prompt_cache_hit_tokens
diff --git a/internal/lua/adapters/kimicode.lua b/internal/lua/adapters/kimicode.lua
index 2d6d4fa..9d10c54 100644
--- a/internal/lua/adapters/kimicode.lua
+++ b/internal/lua/adapters/kimicode.lua
@@ -65,7 +65,7 @@ function adapter.transform_response(raw_body)
         unified.token_usage.completion = resp.usage.completion_tokens or 0
         unified.token_usage.total = resp.usage.total_tokens or 0
         local hit = 0
-        if type(resp.usage.prompt_tokens_details) == "table" and (resp.usage.prompt_tokens_details.cached_tokens or 0) > 0 then
+        if type(resp.usage.prompt_tokens_details) == "table" and resp.usage.prompt_tokens_details.cached_tokens ~= nil then
             hit = resp.usage.prompt_tokens_details.cached_tokens
             unified.token_usage.prompt_tokens_details = { cached_tokens = hit }
         end
@@ -118,7 +118,7 @@ function adapter.transform_stream_chunk(raw_chunk)
             completion = chunk.usage.completion_tokens or chunk.usage.completion or 0,
             total = chunk.usage.total_tokens or chunk.usage.total or 0,
         }
-        if type(chunk.usage.prompt_tokens_details) == "table" and (chunk.usage.prompt_tokens_details.cached_tokens or 0) > 0 then
+        if type(chunk.usage.prompt_tokens_details) == "table" and chunk.usage.prompt_tokens_details.cached_tokens ~= nil then
             uses.prompt_tokens_details = { cached_tokens = chunk.usage.prompt_tokens_details.cached_tokens }
         elseif (chunk.usage.prompt_cache_hit_tokens or 0) > 0 then
             uses.prompt_cache_hit_tokens = chunk.usage.prompt_cache_hit_tokens
diff --git a/internal/lua/adapters/mistral.lua b/internal/lua/adapters/mistral.lua
index e33264f..75004b7 100644
--- a/internal/lua/adapters/mistral.lua
+++ b/internal/lua/adapters/mistral.lua
@@ -33,7 +33,7 @@ function adapter.transform_response(raw_body)
         unified.token_usage.completion = resp.usage.completion_tokens or 0
         unified.token_usage.total = resp.usage.total_tokens or 0
         local hit = 0
-        if type(resp.usage.prompt_tokens_details) == "table" and (resp.usage.prompt_tokens_details.cached_tokens or 0) > 0 then
+        if type(resp.usage.prompt_tokens_details) == "table" and resp.usage.prompt_tokens_details.cached_tokens ~= nil then
             hit = resp.usage.prompt_tokens_details.cached_tokens
             unified.token_usage.prompt_tokens_details = { cached_tokens = hit }
         end
@@ -85,7 +85,7 @@ function adapter.transform_stream_chunk(raw_chunk)
             completion = chunk.usage.completion_tokens or chunk.usage.completion or 0,
             total = chunk.usage.total_tokens or chunk.usage.total or 0,
         }
-        if type(chunk.usage.prompt_tokens_details) == "table" and (chunk.usage.prompt_tokens_details.cached_tokens or 0) > 0 then
+        if type(chunk.usage.prompt_tokens_details) == "table" and chunk.usage.prompt_tokens_details.cached_tokens ~= nil then
             uses.prompt_tokens_details = { cached_tokens = chunk.usage.prompt_tokens_details.cached_tokens }
         elseif (chunk.usage.prompt_cache_hit_tokens or 0) > 0 then
             uses.prompt_cache_hit_tokens = chunk.usage.prompt_cache_hit_tokens
diff --git a/internal/lua/adapters/openai.lua b/internal/lua/adapters/openai.lua
index 83130a4..5538df5 100644
--- a/internal/lua/adapters/openai.lua
+++ b/internal/lua/adapters/openai.lua
@@ -38,7 +38,7 @@ function adapter.transform_response(raw_body)
         -- prompt_tokens_details.cached_tokens (falls back to the legacy
         -- standalone field), so both shapes reach clients.
         local hit = 0
-        if type(resp.usage.prompt_tokens_details) == "table" and (resp.usage.prompt_tokens_details.cached_tokens or 0) > 0 then
+        if type(resp.usage.prompt_tokens_details) == "table" and resp.usage.prompt_tokens_details.cached_tokens ~= nil then
             hit = resp.usage.prompt_tokens_details.cached_tokens
             unified.token_usage.prompt_tokens_details = { cached_tokens = hit }
         end
@@ -93,7 +93,7 @@ function adapter.transform_stream_chunk(raw_chunk)
             completion = chunk.usage.completion_tokens or chunk.usage.completion or 0,
             total = chunk.usage.total_tokens or chunk.usage.total or 0,
         }
-        if type(chunk.usage.prompt_tokens_details) == "table" and (chunk.usage.prompt_tokens_details.cached_tokens or 0) > 0 then
+        if type(chunk.usage.prompt_tokens_details) == "table" and chunk.usage.prompt_tokens_details.cached_tokens ~= nil then
             uses.prompt_tokens_details = { cached_tokens = chunk.usage.prompt_tokens_details.cached_tokens }
         elseif (chunk.usage.prompt_cache_hit_tokens or 0) > 0 then
             uses.prompt_cache_hit_tokens = chunk.usage.prompt_cache_hit_tokens
diff --git a/internal/lua/adapters/opencode.lua b/internal/lua/adapters/opencode.lua
index 80acb83..63c07df 100644
--- a/internal/lua/adapters/opencode.lua
+++ b/internal/lua/adapters/opencode.lua
@@ -101,7 +101,7 @@ function adapter.transform_response(raw_body)
         unified.token_usage.completion = resp.usage.completion_tokens or 0
         unified.token_usage.total = resp.usage.total_tokens or 0
         local hit = 0
-        if type(resp.usage.prompt_tokens_details) == "table" and (resp.usage.prompt_tokens_details.cached_tokens or 0) > 0 then
+        if type(resp.usage.prompt_tokens_details) == "table" and resp.usage.prompt_tokens_details.cached_tokens ~= nil then
             hit = resp.usage.prompt_tokens_details.cached_tokens
             unified.token_usage.prompt_tokens_details = { cached_tokens = hit }
         end
@@ -158,7 +158,7 @@ function adapter.transform_stream_chunk(raw_chunk)
             completion = chunk.usage.completion_tokens or chunk.usage.completion or 0,
             total = chunk.usage.total_tokens or chunk.usage.total or 0,
         }
-        if type(chunk.usage.prompt_tokens_details) == "table" and (chunk.usage.prompt_tokens_details.cached_tokens or 0) > 0 then
+        if type(chunk.usage.prompt_tokens_details) == "table" and chunk.usage.prompt_tokens_details.cached_tokens ~= nil then
             uses.prompt_tokens_details = { cached_tokens = chunk.usage.prompt_tokens_details.cached_tokens }
         elseif (chunk.usage.prompt_cache_hit_tokens or 0) > 0 then
             uses.prompt_cache_hit_tokens = chunk.usage.prompt_cache_hit_tokens
diff --git a/internal/lua/adapters/sensenova.lua b/internal/lua/adapters/sensenova.lua
index db1d2ad..ed37d60 100644
--- a/internal/lua/adapters/sensenova.lua
+++ b/internal/lua/adapters/sensenova.lua
@@ -38,7 +38,7 @@ function adapter.transform_response(raw_body)
         unified.token_usage.completion = resp.usage.completion_tokens or 0
         unified.token_usage.total = resp.usage.total_tokens or 0
         local hit = 0
-        if type(resp.usage.prompt_tokens_details) == "table" and (resp.usage.prompt_tokens_details.cached_tokens or 0) > 0 then
+        if type(resp.usage.prompt_tokens_details) == "table" and resp.usage.prompt_tokens_details.cached_tokens ~= nil then
             hit = resp.usage.prompt_tokens_details.cached_tokens
             unified.token_usage.prompt_tokens_details = { cached_tokens = hit }
         end
@@ -79,7 +79,7 @@ function adapter.transform_stream_chunk(raw_chunk)
             completion = chunk.usage.completion_tokens or chunk.usage.completion or 0,
             total = chunk.usage.total_tokens or chunk.usage.total or 0,
         }
-        if type(chunk.usage.prompt_tokens_details) == "table" and (chunk.usage.prompt_tokens_details.cached_tokens or 0) > 0 then
+        if type(chunk.usage.prompt_tokens_details) == "table" and chunk.usage.prompt_tokens_details.cached_tokens ~= nil then
             uses.prompt_tokens_details = { cached_tokens = chunk.usage.prompt_tokens_details.cached_tokens }
         elseif (chunk.usage.prompt_cache_hit_tokens or 0) > 0 then
             uses.prompt_cache_hit_tokens = chunk.usage.prompt_cache_hit_tokens
diff --git a/internal/types/types.go b/internal/types/types.go
index 34fda89..1f0d7d1 100644
--- a/internal/types/types.go
+++ b/internal/types/types.go
@@ -99,7 +99,9 @@ type TokenUsage struct {
 // PromptTokensDetails is the OpenAI v2 prompt_tokens_details object. Only
 // CachedTokens is emitted (omitempty drops the whole object when zero).
 type PromptTokensDetails struct {
-	CachedTokens int `json:"cached_tokens,omitempty"`
+	// CachedTokens is always emitted (even 0) so clients can distinguish
+	// "upstream reports cache, this request missed" from "no cache data".
+	CachedTokens int `json:"cached_tokens"`
 }
 
 // MarshalJSON emits both the legacy short keys (prompt/completion/total, used