Skip to content

#1 Persistent memory for AI coding agents based on real-world benchmarks

chore(providers): bump Gemini defaults to current GA models (#370)

Rohit Ghumare4b354b7bd855
diff --git a/README.md b/README.md
index 678351a8..7c4dea82 100644
--- a/README.md
+++ b/README.md
@@ -719,7 +719,7 @@ npm install @xenova/transformers
| Provider | Model | Cost | Notes |
|---|---|---|---|
| **Local (recommended)** | `all-MiniLM-L6-v2` | Free | Offline, +8pp recall over BM25-only |
-| Gemini | `text-embedding-004` | Free tier | 1500 RPM |
+| Gemini | `gemini-embedding-001` | Free tier | 100+ languages, 768/1536/3072 dims (MRL), 2048-token input. Replaces `text-embedding-004` ([deprecated, shutdown Jan 14, 2026](https://ai.google.dev/gemini-api/docs/deprecations)) |
| OpenAI | `text-embedding-3-small` | $0.02/1M | Highest quality |
| Voyage AI | `voyage-code-3` | Paid | Optimized for code |
| Cohere | `embed-english-v3.0` | Free trial | General purpose |
diff --git a/src/config.ts b/src/config.ts
index 2898552d..a4b676cf 100644
--- a/src/config.ts
+++ b/src/config.ts
@@ -76,7 +76,7 @@ function detectProvider(env: Record<string, string>): ProviderConfig {
}
return {
provider: "gemini",
- model: env["GEMINI_MODEL"] || "gemini-2.0-flash",
+ model: env["GEMINI_MODEL"] || "gemini-2.5-flash",
maxTokens,
};
}
diff --git a/src/providers/embedding/gemini.ts b/src/providers/embedding/gemini.ts
index 74dca6f5..6cfb2764 100644
--- a/src/providers/embedding/gemini.ts
+++ b/src/providers/embedding/gemini.ts
@@ -2,7 +2,8 @@ import type { EmbeddingProvider } from "../../types.js";
import { getEnvVar } from "../../config.js";
const BATCH_LIMIT = 100;
-const API_BASE = "https://generativelanguage.googleapis.com/v1beta/models/text-embedding-004:batchEmbedContent";
+const MODEL = "models/gemini-embedding-001";
+const API_BASE = `https://generativelanguage.googleapis.com/v1beta/${MODEL}:batchEmbedContents`;
export class GeminiEmbeddingProvider implements EmbeddingProvider {
readonly name = "gemini";
@@ -29,8 +30,9 @@ export class GeminiEmbeddingProvider implements EmbeddingProvider {
headers: { "Content-Type": "application/json" },
body: JSON.stringify({
requests: chunk.map((t) => ({
- model: "models/text-embedding-004",
+ model: MODEL,
content: { parts: [{ text: t }] },
+ outputDimensionality: this.dimensions,
})),
}),
});
@@ -45,10 +47,31 @@ export class GeminiEmbeddingProvider implements EmbeddingProvider {
};
for (const emb of data.embeddings) {
- results.push(new Float32Array(emb.values));
+ results.push(l2Normalize(new Float32Array(emb.values)));
}
}
return results;
}
}
+
+let zeroNormWarned = false;
+
+function l2Normalize(vec: Float32Array): Float32Array {
+ let sum = 0;
+ for (let i = 0; i < vec.length; i++) sum += vec[i]! * vec[i]!;
+ const norm = Math.sqrt(sum);
+ if (norm === 0) {
+ if (!zeroNormWarned) {
+ zeroNormWarned = true;
+ process.stderr.write(
+ `[agentmemory] warn: gemini-embedding-001 returned a zero-norm ` +
+ `embedding (length=${vec.length}); leaving it un-normalized. ` +
+ `Subsequent zero-norm vectors will not be reported.\n`,
+ );
+ }
+ return vec;
+ }
+ for (let i = 0; i < vec.length; i++) vec[i] = vec[i]! / norm;
+ return vec;
+}