Daily vendor refresh (2026-06-12): Mistral Large 3 + Cohere R+ pricing

research_report_2026-06-12.md

@@ -0,0 +1,96 @@
+## 0. Diff & Freshness
+
+- **Report date**: 2026-06-12
+- **Prompt version**: v1
+- **Prior catalog**: present in-repo (`src/data/*.ts`) as of main branch HEAD on 2026-06-12.
+
+### Delta (material)
+
+- **Price changes**
+  - `mistral-large-3`: input/output **$2.00/$6.00 → $0.50/$1.50** USD/Mtok (Mistral pricing page table).
+  - `command-r-plus-2025`: input/output **$3.00/$15.00 → $2.50/$10.00** USD/Mtok (Cohere pricing page, “Command R+ 08-2024”).
+
+- **No new providers** (locked list honored).
+- **No verified new model IDs** found today for the locked provider list beyond what is already in repo.
+
+---
+
+# 1. Models (Model)
+
+## Anthropic
+- No verified new GA model IDs found on the pricing page today; pricing page lists Opus 4.5–4.8, Sonnet 4–4.6, Haiku 4.5, plus “Claude Fable 5” and “Claude Mythos 5 (limited availability)” (not currently represented in repo; unclear GA status). [1]
+
+References:
+1. https://docs.anthropic.com/en/docs/about-claude/pricing
+
+## OpenAI
+- No verified new model IDs beyond what is already in repo; the OpenAI API pricing page currently lists GPT-5.5, GPT-5.4, GPT-5.4 mini, plus “GPT-Realtime-2” variants and “GPT-Image-2” entries. [1]
+
+References:
+1. https://openai.com/api/pricing/
+
+## Google
+- No new model IDs beyond what is already in repo are required for material pricing changes today; Google’s Gemini API pricing page includes Gemini 2.5 Pro, 2.5 Flash, 2.5 Flash-Lite, Gemini 3 Flash Preview, plus additional preview IDs. [1]
+
+References:
+1. https://ai.google.dev/gemini-api/docs/pricing
+
+## Meta
+- No update performed today (no primary-source change retrieved).
+
+## Mistral
+- Pricing page lists “Mistral Large 3” with API price $0.50 input / $1.50 output per 1M tokens (model already exists in repo). [1]
+
+References:
+1. https://docs.mistral.ai/getting-started/pricing/
+
+## DeepSeek
+- DeepSeek API pricing table lists `deepseek-v4-flash` and `deepseek-v4-pro` with 1M context length and max output 384K (these models already exist in repo). [1]
+
+References:
+1. https://api-docs.deepseek.com/quick_start/pricing
+
+## xAI
+- xAI “Models” page lists Grok 4.3 and Grok Build 0.1 with pricing and context windows (repo currently tracks Grok 4.3 only). [1]
+
+References:
+1. https://docs.x.ai/developers/models
+
+## Cohere
+- Cohere pricing page lists “Command R+ 08-2024” at $2.50 input / $10.00 output per 1M tokens (repo currently tracks a single `command-r-plus-2025` entry). [1]
+
+References:
+1. https://cohere.com/pricing
+
+---
+
+# 2. Pricing (Pricing)
+
+## Verified pricing rows (USD per 1M tokens)
+
+| modelId | inputPerMTokUsd | outputPerMTokUsd | cachedInputPerMTokUsd | batchInputPerMTokUsd | notes | Evidence |
+|---|---:|---:|---:|---:|---|---|
+| mistral-large-3 | 0.5 | 1.5 |  |  | Mistral pricing page also notes batch processing gets a 50% discount (no explicit per-model batch row). | [Mistral pricing page](https://docs.mistral.ai/getting-started/pricing/) |
+| command-r-plus-2025 | 2.5 | 10.0 |  |  | Cohere page lists “Command R+ 08-2024” at $2.50/$10.00 (existing customers). | [Cohere pricing page](https://cohere.com/pricing) |
+
+References:
+1. https://docs.mistral.ai/getting-started/pricing/
+2. https://cohere.com/pricing
+
+---
+
+# 3. Benchmarks (Benchmark)
+
+- No benchmark refresh performed today (only pricing deltas were material/verified).
+
+---
+
+# 4. Use-cases (UseCase)
+
+- No use-case refresh performed today.
+
+---
+
+# 5. Historic-usage (HistoricUsageCase)
+
+- No historic-usage refresh performed today.

src/data/pricing.ts

@@ -95,9 +95,10 @@ export const pricing: Pricing[] = [
   },
   {
     modelId: "mistral-large-3",
-    inputPerMTokUsd: 2.0,
-    outputPerMTokUsd: 6.0,
-    batchInputPerMTokUsd: 1.0,
+    inputPerMTokUsd: 0.5,
+    outputPerMTokUsd: 1.5,
+    notes:
+      "Mistral pricing page lists Mistral Large 3 at $0.50 input / $1.50 output per 1M tokens; page also notes batch processing gets a 50% discount.",
   },
     {
     modelId: "deepseek-v4-flash",
@@ -115,8 +116,9 @@ export const pricing: Pricing[] = [
   },
     {
     modelId: "command-r-plus-2025",
-    inputPerMTokUsd: 3.0,
-    outputPerMTokUsd: 15.0,
-    notes: "Cohere Command R+ pricing (Cohere pricing page).",
+    inputPerMTokUsd: 2.5,
+    outputPerMTokUsd: 10.0,
+    notes:
+      "Cohere pricing page lists Command R+ 08-2024 at $2.50 input / $10.00 output per 1M tokens (existing customers).",
   },
 ];