copilot-custom-endpoint 2.1.6 → 2.1.8
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +54 -39
- package/docs/example-config.md +18 -2
- package/docs/models/glm.md +33 -14
- package/docs/pricing.md +17 -5
- package/package.json +1 -1
package/README.md
CHANGED
|
@@ -29,6 +29,7 @@ That's it. No code, no servers to manage (unless the model specifically needs th
|
|
|
29
29
|
| **Qwen 3.7 Max** | DashScope | Optional (recommended) | ❌ | [Manual setup](docs/models/qwen.md) |
|
|
30
30
|
| **Qwen 3.8 Max** | DashScope | Optional (recommended) | ✅ | [Manual setup](docs/models/qwen.md) |
|
|
31
31
|
| **MiniMax M3** | MiniMax | No | ✅ | [Extension](https://github.com/tugudush/minimax-copilot) ★ / [Manual setup](docs/models/minimax.md) |
|
|
32
|
+
| **GLM 5.3 Flash** | Z.ai | No | ✅ | [Manual setup](docs/models/glm.md) |
|
|
32
33
|
| **GLM 5.3 / 5.2 / 5.1** | Z.ai | No | ✅ via extension proxy | [Manual setup](docs/models/glm.md) |
|
|
33
34
|
| **GLM 5V Turbo** | Z.ai | No | ✅ | [Manual setup](docs/models/glm.md) |
|
|
34
35
|
| **DeepSeek V4 Flash 0731** | DeepSeek | No (uses an extension) | ❌ | [Extension](https://marketplace.visualstudio.com/items?itemName=Vizards.deepseek-v4-for-copilot) / [setup notes](docs/models/deepseek.md) |
|
|
@@ -108,43 +109,47 @@ npx copilot-custom-endpoint clean # Remove debug_log/
|
|
|
108
109
|
|
|
109
110
|
All prices are **USD per 1M tokens** (non-cached). 1 AI credit = $0.01. To convert to AI credits, multiply by 100 (e.g., $5.00/1M = 500 credits/1M). Session cost assumes ~10K input + ~2K output tokens per turn, 50 turns. DeepSeek V4 estimates use peak rates; official off-peak rates are half.
|
|
110
111
|
|
|
111
|
-
| Model
|
|
112
|
-
|
|
|
113
|
-
| **
|
|
114
|
-
| **
|
|
115
|
-
| **
|
|
116
|
-
| **
|
|
117
|
-
| **
|
|
118
|
-
| **
|
|
119
|
-
| **
|
|
120
|
-
| **
|
|
121
|
-
| **
|
|
122
|
-
| **Gemini 3.
|
|
123
|
-
| **
|
|
124
|
-
| **
|
|
125
|
-
| **
|
|
126
|
-
| **
|
|
127
|
-
| **
|
|
128
|
-
| **
|
|
129
|
-
| **
|
|
130
|
-
| **
|
|
131
|
-
| **Qwen 3.
|
|
132
|
-
| **
|
|
133
|
-
| **
|
|
134
|
-
| **
|
|
135
|
-
| **
|
|
136
|
-
| **
|
|
137
|
-
| **Gemini 3.
|
|
138
|
-
| **
|
|
139
|
-
| **
|
|
140
|
-
| **
|
|
141
|
-
| **
|
|
142
|
-
| **Claude
|
|
143
|
-
| **Claude
|
|
144
|
-
| **
|
|
145
|
-
| **Claude Opus 4.
|
|
146
|
-
| **GPT-5.
|
|
147
|
-
| **Claude
|
|
112
|
+
| Model | Provider | Cost per intelligence | Intelligence Score | Est. session | Vision | Context window |
|
|
113
|
+
| ----------------------------- | --------- | --------------------- | ------------------ | ------------ | ------ | -------------- |
|
|
114
|
+
| **GLM 5.3 Flash** ¹⁸ | Z.ai | **~$0.0022** | **57.5** | ~$0.13 | ✅ | 1M |
|
|
115
|
+
| **MiMo V2.5** | Xiaomi | **~$0.0026** | **38.0** | ~$0.10 | ✅ | 1M |
|
|
116
|
+
| **GPT-5.6 Luna** | OpenAI | **~$0.0042** | **52.3** | ~$0.22 | ✅ | 1M |
|
|
117
|
+
| **MiniMax M3** | MiniMax | **~$0.0060** | **45.4** | ~$0.27 | ✅ | 1M |
|
|
118
|
+
| **DeepSeek V4 Flash 0731** | DeepSeek | **~$0.0068** | **51.8** | ~$0.35 | ❌ | 1M |
|
|
119
|
+
| **MiMo V2.5 Pro** | Xiaomi | **~$0.0070** | **42.9** | ~$0.30 | ❌ | 1M |
|
|
120
|
+
| **Qwen 3.7 Plus** ¹² | DashScope | **~$0.0074** | **39.4** | ~$0.29 | ✅ | 1M |
|
|
121
|
+
| **MiniMax M3 Priority**⁴ | MiniMax | **~$0.0090** | **45.4** | ~$0.41 | ✅ | 1M |
|
|
122
|
+
| **Gemini 3.8 Flash** ¹⁹ | Google | **~$0.013** | **58.7** | ~$0.75 | ✅ | 1M |
|
|
123
|
+
| **Gemini 3.7 Flash** ¹³ | Google | **~$0.013** | **56** | ~$0.75 | ✅ | 1M |
|
|
124
|
+
| **Qwen 3.6 Plus** ¹⁶ | DashScope | **~$0.014** | **40.0** | ~$0.55 | ✅ | 1M |
|
|
125
|
+
| **Gemini 3.6 Flash** ⁸ | Google | **~$0.015** | **51.6** | ~$0.75 | ✅ | 1M |
|
|
126
|
+
| **GLM 5.3** ¹⁷ | Z.ai | **~$0.019** | **59.5** | ~$1.14 | ❌ | 1M |
|
|
127
|
+
| **Kimi K2.6** | Moonshot | **~$0.020** | **45.0** | ~$0.88 | ✅ | 262K |
|
|
128
|
+
| **DeepSeek V4 Pro 0813** | DeepSeek | **~$0.020** | **53.2** | ~$1.06 | ❌ | 1M |
|
|
129
|
+
| **GPT-5.4 mini** | OpenAI | **~$0.020** | **40.9** | ~$0.83 | ❌ | 400K |
|
|
130
|
+
| **Kimi K2.7 Code** | Moonshot | **~$0.020** | **43.0** | ~$0.88 | ✅ | 262K |
|
|
131
|
+
| **GLM 5.2** | Z.ai | **~$0.022** | **52.6** | ~$1.14 | ❌ | 1M |
|
|
132
|
+
| **Qwen 3.7 Max** ¹² | DashScope | **~$0.025** | **46.7** | ~$1.18 | ❌ | 1M |
|
|
133
|
+
| **Grok 4.6** ¹⁴ | xAI | **~$0.026** | **61** | ~$1.60 | ✅ | 500K |
|
|
134
|
+
| **Qwen 3.8 Max** ¹⁰ | DashScope | **~$0.028** | **58.1** | ~$1.60 | ✅ | 1M |
|
|
135
|
+
| **GLM 5.1** | Z.ai | **~$0.028** | **41.0** | ~$1.14 | ❌ | 200K |
|
|
136
|
+
| **GLM 5V Turbo** | Z.ai | **~$0.029** | **35.0** ³ | ~$1.00 | ✅ | 200K |
|
|
137
|
+
| **Grok 4.5** ¹¹ | xAI | **~$0.029** | **55.8** | ~$1.60 | ✅ | 500K |
|
|
138
|
+
| **Gemini 3.5 Flash** | Google | **~$0.032** | **52.0** | ~$1.65 | ✅ | 1M |
|
|
139
|
+
| **GPT-5.6 Terra** | OpenAI | **~$0.039** | **56.6** | ~$2.20 | ✅ | 1M |
|
|
140
|
+
| **Gemini 3.1 Pro** | Google | **~$0.046** | **47.7** | ~$2.20 | ✅ | 1M |
|
|
141
|
+
| **Kimi K3** ⁶ | Moonshot | **~$0.050** | **59.7** | ~$3.00 | ✅ | 1M |
|
|
142
|
+
| **GPT-5.4** | OpenAI | **~$0.052** | **53.1** | ~$2.75 | ✅ | 1M |
|
|
143
|
+
| **Claude Sonnet 5** ⁵ | Anthropic | **~$0.054** | **55.3** | ~$3.00 | ✅ | 1M |
|
|
144
|
+
| **Claude Sonnet 4.6** | Anthropic | **~$0.062** | **48.4** | ~$3.00 | ✅ | 1M |
|
|
145
|
+
| **Claude Opus 5** ⁹ | Anthropic | **~$0.079** | **63.1** | ~$5.00 | ✅ | 1M |
|
|
146
|
+
| **Claude Opus 4.8** | Anthropic | **~$0.087** | **57.3** | ~$5.00 | ✅ | 1M |
|
|
147
|
+
| **GPT-5.6 Sol** | OpenAI | **~$0.090** | **60.9** | ~$5.50 | ✅ | 1M |
|
|
148
|
+
| **Claude Opus 4.7** | Anthropic | **~$0.091** | **55.0** | ~$5.00 | ✅ | 1M |
|
|
149
|
+
| **GPT-5.5** | OpenAI | **~$0.098** | **56.3** | ~$5.50 | ✅ | 1M |
|
|
150
|
+
| **Claude Fable 5.1** ²¹ | Anthropic | **~$0.15** | **65.7** | ~$10.00 | ✅ | 1M |
|
|
151
|
+
| **Claude Fable 5** ¹⁵ | Anthropic | **~$0.161** | **62.1** | ~$10.00 | ✅ | 1M |
|
|
152
|
+
| **Gemini 3.8 Flash Cyber** ²⁰ | Google | — | — | — | — | — |
|
|
148
153
|
|
|
149
154
|
¹ Gemini 3.1 Pro pricing applies to prompts ≤200K tokens.
|
|
150
155
|
|
|
@@ -172,12 +177,20 @@ All prices are **USD per 1M tokens** (non-cached). 1 AI credit = $0.01. To conve
|
|
|
172
177
|
|
|
173
178
|
¹⁴ **Grok 4.6** (xAI, released August 12, 2026) is now a **GitHub Copilot native** model (GA). AA Intelligence Index **61** (high preset, #6/188). 500K context, text + image input. Priced at $2.00 / $0.50 / $6.00 per 1M input/cached/output. See the [AA model page](https://artificialanalysis.ai/models/grok-4-6), the [xAI announcement](https://x.ai/news/grok-4-6), and the [GitHub Copilot models & pricing page](https://docs.github.com/en/copilot/reference/copilot-billing/models-and-pricing).
|
|
174
179
|
|
|
175
|
-
¹⁵ **Claude Fable 5** (Anthropic, released June 9, 2026) is now a **GitHub Copilot native** model (GA, Powerful). AA Intelligence Index **62.1** (#3/188) confirmed by [Artificial Analysis](https://artificialanalysis.ai/models/claude-fable-5). **#1 in all four Arena leaderboards** (Text, Agent, Code, Overall). 1M context, text + image input, adaptive reasoning. Priced at $10.00 / $50.00 per MTok input/output ($1.00 cached, 90% cache discount). Uses the newer Claude tokenizer (~30% more tokens than pre-4.7 models). See the [GitHub Copilot models & pricing page](https://docs.github.com/en/copilot/reference/copilot-billing/models-and-pricing).
|
|
180
|
+
¹⁵ **Claude Fable 5** (Anthropic, released June 9, 2026) is now a **GitHub Copilot native** model (GA, Powerful). AA Intelligence Index **62.1** (#3/188) confirmed by [Artificial Analysis](https://artificialanalysis.ai/models/claude-fable-5). **#1 in all four Arena leaderboards** (Text, Agent, Code, Overall). 1M context, text + image input, adaptive reasoning. Priced at $10.00 / $50.00 per MTok input/output ($1.00 cached, 90% cache discount). Uses the newer Claude tokenizer (~30% more tokens than pre-4.7 models). **Superseded as Anthropic's flagship by Claude Fable 5.1 (September 2026 — see footnote ²¹)**. See the [GitHub Copilot models & pricing page](https://docs.github.com/en/copilot/reference/copilot-billing/models-and-pricing).
|
|
176
181
|
|
|
177
182
|
¹⁶ **Qwen 3.6 Plus** (DashScope, released April 2, 2026) — AA Intelligence Index **40.0**. Priced at $0.50 / $3.00 per MTok input/output ($0.05 cached). 1M context, text + image + video input. **Deprecated** in favor of Qwen 3.7 Plus.
|
|
178
183
|
|
|
179
184
|
¹⁷ **GLM 5.3** (Z.ai, released August 18, 2026) — new flagship. AA Intelligence Index **59.5** (OpenRouter AA-sourced; AA page lists rounded **60**, #8/182). Priced identically to GLM 5.2: $1.40 / $0.26 / $4.40 per 1M input/cached/output. 1M context, text-only, always-thinking with `reasoning_effort` (low/high/max, default max). Released too recently for Arena rankings. See the [AA model page](https://artificialanalysis.ai/models/glm-5-3) and [Z.ai docs](https://docs.z.ai/guides/llm/glm-5.3).
|
|
180
185
|
|
|
186
|
+
¹⁸ **GLM 5.3 Flash** (Z.ai, released August 26, 2026) — Z.ai's first native multimodal GLM-5 model (text + image input, text output). Hybrid sparse + linear attention architecture, 320B total / 18B active parameters (open weights, MIT), 1M context window. AA Intelligence Index **57.5** (Coding **71.5**, Agentic **58.2**). Priced at **$0.15 / $0.03 / $0.50** per 1M input/cached/output (list; 50% off launch promo through September 9, 2026). Always-thinking with `reasoning_effort` (low/high/max, default max). The cheapest cost-per-intelligence row in this table (~$0.0022, ~$0.13/session). See [docs/models/glm.md](docs/models/glm.md) and the [AA model page](https://artificialanalysis.ai/models/glm-5-3-flash).
|
|
187
|
+
|
|
188
|
+
¹⁹ **Gemini 3.8 Flash** (`gemini-3.8-flash`, released September 2, 2026) — Google's most intelligent Flash model, built for long-horizon software engineering and autonomous agents. AA Intelligence Index **58.7** (OpenRouter AA-sourced; AA page rounds to **59**, #17/196), Coding **76.3**, Agentic **50.0**. 1M context, text + image + speech + video + PDF input. Configurable thinking (low/medium/high). **Not yet GitHub Copilot native** — priced from the Gemini API at the same promotional rates as 3.7 Flash ($0.75 / $3.75 per MTok input/output, $0.075 cached) through December 31, 2026, then $1.50 / $7.50. Fast (305.5 t/s) but very verbose — budget more output tokens than the score alone suggests. See the [Google announcement](https://blog.google/innovation-and-ai/models-and-research/gemini-models/3-8-flash-and-3-8-flash-cyber/) and the [AA model page](https://artificialanalysis.ai/models/gemini-3-8-flash).
|
|
189
|
+
|
|
190
|
+
²⁰ **Gemini 3.8 Flash Cyber** (released September 2, 2026) — cybersecurity-specialized variant of 3.8 Flash (autonomous vulnerability discovery + automated patching). **Restricted access** via Google's [Fairwind Program](https://deepmind.google/fairwind-program/) — no public API pricing and no AA Intelligence Index run, hence the `—` cells. Vendor-reported CyberGym Pass@1 **86.2** and CWE-Bench Pass@1 **47.2**. See the [Google announcement](https://blog.google/innovation-and-ai/models-and-research/gemini-models/3-8-flash-and-3-8-flash-cyber/).
|
|
191
|
+
|
|
192
|
+
²¹ **Claude Fable 5.1** (released September 1, 2026) — Anthropic's new flagship; AA Intelligence Index **65.7** (OpenRouter AA-sourced; AA page rounds to **66**, #1/196), Coding **81.6**, Agentic **61.3** — now the top-scoring model in this table. GitHub Copilot native (GA, Powerful). 1M context, text + image input, adaptive reasoning. Priced at $10.00 / $50.00 per MTok input/output ($0.25 cached — down from Fable 5's $1.00; 97.5% cache discount; $12.50 cache write). Uses the newer Claude tokenizer (~30% more tokens than pre-4.7 models). Too new for Arena rankings. See the [GitHub Copilot models & pricing page](https://docs.github.com/en/copilot/reference/copilot-billing/models-and-pricing) and the [AA model page](https://artificialanalysis.ai/models/claude-fable-5-1).
|
|
193
|
+
|
|
181
194
|
For footnotes, sources, and detailed notes (cache behavior, tiered pricing, free quotas) see [docs/pricing.md](docs/pricing.md). For a copy-paste config containing **all providers at once**, see [docs/example-config.md](docs/example-config.md).
|
|
182
195
|
|
|
183
196
|
> **👤 Personal picks** —
|
|
@@ -187,14 +200,16 @@ For footnotes, sources, and detailed notes (cache behavior, tiered pricing, free
|
|
|
187
200
|
> - **GPT-5.6 Luna** (Copilot native) — scores **52.3**, ~$0.22/session, vision-capable. It is cheaper than MiniMax M3 (45.4) and consumes fewer AI credits after OpenAI's July price cut.
|
|
188
201
|
> - **DeepSeek V4 Flash 0731** (extension) — scores **51.8** after the AA v4.1.1 benchmark refresh, ~$0.35/session at peak rates (~$0.18 off-peak). It is the lowest-cost text-only model above 50 in this table; pair it with a vision model when you need images.
|
|
189
202
|
> - **Gemini 3.7 Flash** (Copilot native, new August 13) — scores **56** at ~$0.75/session under promotional pricing. The best-value Google model in Copilot: it beats Gemini 3.5 Flash (52.0) and Grok 4.5 (55.8) on intelligence at roughly half the session cost, with 1M context, full multimodal input, and AA's fastest output speed. A strong upgrade pick if you use Google models.
|
|
203
|
+
> - **Gemini 3.8 Flash** (Gemini API, new September 2) — scores **58.7** at the same ~$0.75/session promotional price as 3.7 Flash, with +2.7 intelligence, 1M context, and full multimodal input. **Not yet Copilot native**, so it's a comparison row here rather than a config — and it runs very verbose (AA measured 120M output tokens on the Intelligence Index), so budget more output tokens than the score alone suggests.
|
|
190
204
|
>
|
|
191
205
|
> **MiniMax M3** (45.4, ~$0.27/session) is still the best **direct** custom-endpoint option — no proxy or extension, vision, 1M context, strong coding. Pick M3 when you want a plug-and-play custom endpoint with vision.
|
|
192
206
|
>
|
|
193
|
-
> For **plan mode / architecture & design thinking**, **Qwen 3.8 Max**, **Kimi K3**, **GLM 5.3**, **GPT-5.6 Luna**, and **DeepSeek V4 Flash 0731** are worth considering:
|
|
207
|
+
> For **plan mode / architecture & design thinking**, **Qwen 3.8 Max**, **Kimi K3**, **GLM 5.3**, **GLM 5.3 Flash**, **GPT-5.6 Luna**, and **DeepSeek V4 Flash 0731** are worth considering:
|
|
194
208
|
>
|
|
195
209
|
> - **Qwen 3.8 Max** — scores **58.1** (#5 overall, the #2 custom-endpoint model) — the default **multimodal** planning pick. Near-flagship intelligence at ~$1.60/session, with 1M context and vision. Strong for whole-codebase planning, architecture, and large spec review.
|
|
196
210
|
> - **Kimi K3** — new flagship. Highest intelligence score among custom-endpoint models (**59.7**, #4 overall). Best for complex reasoning, long-horizon planning, and architecture decisions when you want the absolute best — but at ~$3.00/session it's ~2× Qwen 3.8 Max for +1.6 points, so **reserve it for hard problems**.
|
|
197
211
|
> - **GLM 5.3** — new flagship, scores **59.5** (now the #2 custom-endpoint model, just behind Kimi K3's 59.7) at ~$1.14/session. 1M context, text-only, always-thinking. It is the strongest text-only planning value above 59 and offers best-in-class agentic coding per Z.ai (50% gain over 5.2 on their Code Bench) — ideal for whole-codebase planning and large spec review. ~4× MiniMax M3, so **not a daily driver for agent mode** — reserve it for planning, then hand implementation off to a cheaper model.
|
|
212
|
+
> - **GLM 5.3 Flash** — new (Aug 26): scores **57.5** at just ~$0.13/session (list) — the **cheapest 57+ model** in the table and the first native multimodal GLM-5. Near-GLM-5.3 agentic coding (Z.ai: DeepSWE v1.1 63.4 vs 46.2) at ~9× less cost, with vision. A strong daily-driver planning + image-review pick until the 50% promo ends.
|
|
198
213
|
> - **GPT-5.6 Luna** (Copilot native) — scores **52.3** and is now the cheapest 50+ model. OpenAI cut its API price to $0.20/$1.20 per 1M input/output tokens; in Copilot, the equivalent benefit is lower AI-credit consumption, while your subscription price and monthly allowance stay unchanged. Ideal for high-quality planning on a budget.
|
|
199
214
|
> - **DeepSeek V4 Flash 0731** (extension) — scores **51.8** at ~$0.35/session at peak rates (~$0.18 off-peak) — the lower-cost planning pick. Text-only, so pair it with a vision model for design review, but it remains strong for reasoning-heavy planning.
|
|
200
215
|
>
|
package/docs/example-config.md
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
# Full example config
|
|
2
2
|
|
|
3
|
-
Here's a complete, real-world `chatLanguageModels.json` that combines **the `customendpoint` providers from the live `chatLanguageModels.json`**, plus the additional GLM models (`glm-5.3`, `glm-5.2`, `glm-5.1`) that are validated in [docs/models/glm.md](models/glm.md) but not present in every live config. Copy what you need, leave the rest out.
|
|
3
|
+
Here's a complete, real-world `chatLanguageModels.json` that combines **the `customendpoint` providers from the live `chatLanguageModels.json`**, plus the additional GLM models (`glm-5.3-flash`, `glm-5.3`, `glm-5.2`, `glm-5.1`) that are validated in [docs/models/glm.md](models/glm.md) but not present in every live config. Copy what you need, leave the rest out.
|
|
4
4
|
|
|
5
5
|
> **⚠️ Reminder:** After adding models, you must also configure the **Utility Small Model** setting in VS Code. Open Settings → search **"Chat: Utility Small Model"** → pick your fastest model (e.g., DeepSeek V4 Flash or MiMo V2.5). Without this, custom-endpoint models may not function correctly. See the [main README](../README.md#4-configure-the-utility-small-model) for details.
|
|
6
6
|
>
|
|
@@ -163,6 +163,22 @@ Here's a complete, real-world `chatLanguageModels.json` that combines **the `cus
|
|
|
163
163
|
"apiKey": "",
|
|
164
164
|
"apiType": "chat-completions",
|
|
165
165
|
"models": [
|
|
166
|
+
{
|
|
167
|
+
"id": "glm-5.3-flash",
|
|
168
|
+
"name": "GLM 5.3 Flash (multimodal)",
|
|
169
|
+
"url": "https://api.z.ai/api/paas/v4/chat/completions",
|
|
170
|
+
"toolCalling": true,
|
|
171
|
+
"vision": true,
|
|
172
|
+
"streaming": true,
|
|
173
|
+
"maxInputTokens": 1048576,
|
|
174
|
+
"maxOutputTokens": 131072,
|
|
175
|
+
"requestBody": {
|
|
176
|
+
"thinking": { "type": "enabled" },
|
|
177
|
+
"reasoning_effort": "max",
|
|
178
|
+
"temperature": 1.0,
|
|
179
|
+
"top_p": 0.95
|
|
180
|
+
}
|
|
181
|
+
},
|
|
166
182
|
{
|
|
167
183
|
"id": "glm-5.3",
|
|
168
184
|
"name": "GLM 5.3 (text)",
|
|
@@ -237,6 +253,6 @@ If you only need one provider, jump straight to its setup guide:
|
|
|
237
253
|
- [Qwen 3.7 Plus / 3.7 Max](qwen.md)
|
|
238
254
|
- [Xiaomi MiMo (V2.5 / V2.5 Pro)](mimo.md)
|
|
239
255
|
- [MiniMax M3 (Standard)](minimax.md) and [MiniMax M3 (Priority)](minimax.md#5-m3-priority-tier-optional)
|
|
240
|
-
- [GLM (5.3 / 5.2 / 5.1 / 5V Turbo)](glm.md)
|
|
256
|
+
- [GLM (5.3 Flash / 5.3 / 5.2 / 5.1 / 5V Turbo)](glm.md)
|
|
241
257
|
|
|
242
258
|
> **DeepSeek V4 Pro 0813 / V4 Flash 0731** use the [DeepSeek V4 for Copilot Chat](https://marketplace.visualstudio.com/items?itemName=Vizards.deepseek-v4-for-copilot) extension. They appear in `chatLanguageModels.json` as `vendor: "deepseek"` (not as a `customendpoint` provider) and are configured via the extension's settings block.
|
package/docs/models/glm.md
CHANGED
|
@@ -18,9 +18,9 @@
|
|
|
18
18
|
| ---------------------- | ------------------------------------------------------------------------------------------------------------------ |
|
|
19
19
|
| Mode | **Direct** (no proxy) |
|
|
20
20
|
| Billing | **Pay-as-You-Go** (PaaS API) — Coding Plan subscription exists but is **not usable** from VS Code custom endpoints |
|
|
21
|
-
| Vision | ✅ Yes (`glm-5v-turbo`
|
|
22
|
-
| Tool calling | ✅ Yes (native multimodal tool use on `glm-5v-turbo`)
|
|
23
|
-
| Context | 1M (`glm-5.3` / `glm-5.2` Solid Lossless Context)
|
|
21
|
+
| Vision | ✅ Yes (`glm-5.3-flash` / `glm-5v-turbo`) |
|
|
22
|
+
| Tool calling | ✅ Yes (native multimodal tool use on `glm-5.3-flash` / `glm-5v-turbo`) |
|
|
23
|
+
| Context | 1M (`glm-5.3-flash` / `glm-5.3` / `glm-5.2` Solid Lossless Context) |
|
|
24
24
|
| Max output | 131072 |
|
|
25
25
|
| Required `requestBody` | `thinking: { type: "enabled" }` + `reasoning_effort: "max"` (recommended) |
|
|
26
26
|
| Endpoint (intl) | `https://api.z.ai/api/paas/v4/chat/completions` |
|
|
@@ -29,12 +29,13 @@
|
|
|
29
29
|
|
|
30
30
|
### Models
|
|
31
31
|
|
|
32
|
-
| Model
|
|
33
|
-
|
|
|
34
|
-
| `glm-5.3`
|
|
35
|
-
| `glm-5.
|
|
36
|
-
| `glm-5.
|
|
37
|
-
| `glm-
|
|
32
|
+
| Model | Vision | Context | Max output | Notes |
|
|
33
|
+
| --------------- | ------ | ------- | ---------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
|
|
34
|
+
| `glm-5.3-flash` | ✅ | 1M | 131072 | **New multimodal Flash** (Aug 26) — 320B/18B hybrid sparse+linear attention; AA **57.5**; always-thinking, `reasoning_effort` `low`/`high`/`max` (default `max`) ($0.15 / $0.50 per 1M) |
|
|
35
|
+
| `glm-5.3` | ❌ | 1M | 131072 | **New flagship** — ~50% coding gain over 5.2 on Z.ai Code Bench; always-thinking, `reasoning_effort` `low`/`high`/`max` (default `max`) ($1.40 / $4.40 per 1M) |
|
|
36
|
+
| `glm-5.2` | ❌ | 1M | 131072 | Previous flagship — "Opus-level" long-context engineering & agentic coding; thinking `enabled` ($1.40 / $4.40 per 1M) |
|
|
37
|
+
| `glm-5.1` | ❌ | 200K | 131072 | Previous flagship — long-horizon / 8h autonomous work; thinking `enabled` ($1.40 / $4.40 per 1M) |
|
|
38
|
+
| `glm-5v-turbo` | ✅ | 200K | 131072 | Multimodal **coding** model — vision-based agentic coding; thinking `enabled` ($1.20 / $4.00 per 1M) |
|
|
38
39
|
|
|
39
40
|
> Other GLM models (`glm-5`, `glm-5-turbo`, `glm-4.5-air`, etc.) are callable on the same endpoint but are intentionally **not** added to the default `chatLanguageModels.json` block below. Add them in the same shape if you need them. Note: `glm-4.6v-flashx` was previously in the default block but has been **removed** because live testing showed it is not reliable for tool calling.
|
|
40
41
|
|
|
@@ -64,6 +65,22 @@
|
|
|
64
65
|
"apiKey": "",
|
|
65
66
|
"apiType": "chat-completions",
|
|
66
67
|
"models": [
|
|
68
|
+
{
|
|
69
|
+
"id": "glm-5.3-flash",
|
|
70
|
+
"name": "GLM 5.3 Flash (multimodal)",
|
|
71
|
+
"url": "https://api.z.ai/api/paas/v4/chat/completions",
|
|
72
|
+
"toolCalling": true,
|
|
73
|
+
"vision": true,
|
|
74
|
+
"streaming": true,
|
|
75
|
+
"maxInputTokens": 1048576,
|
|
76
|
+
"maxOutputTokens": 131072,
|
|
77
|
+
"requestBody": {
|
|
78
|
+
"thinking": { "type": "enabled" },
|
|
79
|
+
"reasoning_effort": "max",
|
|
80
|
+
"temperature": 1.0,
|
|
81
|
+
"top_p": 0.95
|
|
82
|
+
}
|
|
83
|
+
},
|
|
67
84
|
{
|
|
68
85
|
"id": "glm-5.3",
|
|
69
86
|
"name": "GLM 5.3 (text)",
|
|
@@ -150,10 +167,10 @@
|
|
|
150
167
|
|
|
151
168
|
- **Sampling:** recommended `temperature: 1`, `top_p: 0.95`. Z.ai's `temperature` is **hard-capped at `1.0` server-side** — sending `> 1.0` will reject with a 400. Don't set `do_sample: false` from `requestBody` (it would make `temperature`/`top_p` be ignored).
|
|
152
169
|
- **`tool_choice` only supports `auto`** — don't override it (VS Code's default is `auto`).
|
|
153
|
-
- **`clear_thinking` defaults to `true` server-side**, which strips historical `reasoning_content` from prior turns before sending to the model. This is what makes multi-turn tool loops stable. **Do not** set `clear_thinking: false` from `requestBody` — VS Code can't forward `reasoning_content` and the request will fail.
|
|
154
|
-
- **GLM 5.3
|
|
155
|
-
- **Vision** is supported
|
|
156
|
-
- **Context caching** is automatic — `usage.prompt_tokens_details.cached_tokens` reports cache hits; cache writes are currently free.
|
|
170
|
+
- **`clear_thinking` defaults to `true` server-side**, which strips historical `reasoning_content` from prior turns before sending to the model. This is what makes multi-turn tool loops stable. **Do not** set `clear_thinking: false` from `requestBody` — VS Code can't forward `reasoning_content` and the request will fail. (Note: Z.ai's GLM-5.3-Flash docs recommend `thinking.clear_thinking: false` for clients that **do** preserve `reasoning_content` — e.g. Claude Code / agent SDKs — not for the VS Code custom-endpoint path.)
|
|
171
|
+
- **GLM 5.3 and GLM 5.3 Flash are always-thinking** — `thinking.type` only supports `enabled`; disabling reasoning is **not supported**. `reasoning_effort` accepts `low`, `high`, or `max` (default `max`; `max` recommended for coding). If you previously sent `thinking.type: "disabled"` for other GLM models, switch it to `enabled` with `reasoning_effort: "low"` when moving to `glm-5.3` / `glm-5.3-flash`.
|
|
172
|
+
- **Vision** is supported on `glm-5.3-flash` (native multimodal — the first in the GLM-5 series; text + image input, text output) and `glm-5v-turbo` (OpenAI `image_url` content-part format — external URLs and base64 data URIs both work). `glm-5.3`, `glm-5.2` and `glm-5.1` are text-only. For turnkey VS Code video understanding via `glm-5v-turbo`, see [**Video Context MCP**](https://www.videocontextmcp.com/). **Validated 2026-08-28 on `glm-5.3-flash`:** both vision surfaces work through the custom endpoint — chat image attachments (base64 `image_url`) and the `view_image` tool on workspace files. Fine-grained text, layout, and image content were read correctly in both tests.
|
|
173
|
+
- **Context caching** is automatic — `usage.prompt_tokens_details.cached_tokens` reports cache hits; cache writes are currently free. Cache reads for `glm-5.3-flash` are $0.03/M (list; $0.015/M during the launch promo).
|
|
157
174
|
|
|
158
175
|
## Troubleshooting
|
|
159
176
|
|
|
@@ -174,4 +191,6 @@
|
|
|
174
191
|
|
|
175
192
|
All prices are **USD per 1M tokens** (cache miss) on the Z.ai international platform. Per-model input/output rates are listed in the `Cost` column of the [Models](#models) table above.
|
|
176
193
|
|
|
177
|
-
> **Cache writes** are currently **Limited-time Free** for all models. Cached-input pricing is roughly 18% of the input price (e.g. `$1.40` input → `$0.26` cached for `glm-5.3` / `glm-5.2` / `glm-5.1
|
|
194
|
+
> **Cache writes** are currently **Limited-time Free** for all models. Cached-input pricing is roughly 18–20% of the input price (e.g. `$1.40` input → `$0.26` cached for `glm-5.3` / `glm-5.2` / `glm-5.1`; `$0.15` → `$0.03` for `glm-5.3-flash`).
|
|
195
|
+
>
|
|
196
|
+
> **GLM 5.3 Flash** launched **August 26, 2026** at **$0.15 / $0.03 / $0.50** per 1M (input / cached / output) — currently **50% off** (effective **$0.075 / $0.015 / $0.25**) through **September 9, 2026** (24:00 UTC+8). China platform (`bigmodel.cn`) prices in CNY; see the [China pricing page](https://bigmodel.cn/pricing). For the cross-provider comparison, see [docs/pricing.md](../pricing.md).
|
package/docs/pricing.md
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
# Pricing
|
|
2
2
|
|
|
3
|
-
> **Updated:**
|
|
3
|
+
> **Updated:** September 3, 2026 — **Claude Fable 5.1** added (footnote ²³): Anthropic's new flagship (released September 1, 2026), AA Intelligence Index **65.7** — now the **top intelligence score** in the benchmarks table. Same $10 / $50 rates as Fable 5, but cached input cut to $0.25 (97.5% cache discount); it lands just above Fable 5 in the CPI sort (~$0.15 vs ~$0.161), still the second-most expensive scored row. **Gemini 3.8 Flash** added (footnote ²¹): Google's most intelligent Flash model (released September 2, 2026), AA Intelligence Index **58.7**, at the same promotional rates as 3.7 Flash ($0.75 / $0.075 / $3.75 through December 31) — it is **not yet GitHub Copilot native**, so it is priced from the Gemini API. **Gemini 3.8 Flash Cyber** added (footnote ²²): the cybersecurity-specialized sibling, restricted to Google's Fairwind Program — no public API pricing and no AA Intelligence Index run, so its row shows `—` at the bottom of the table. August 28: **GLM 5.3 Flash** added (footnote ²⁰): Z.ai's first native multimodal GLM-5 model (August 26, 2026), AA Intelligence Index **57.5**, launch pricing **$0.15 / $0.03 / $0.50** (50% off through September 9). It is now the **cheapest cost-per-intelligence row** in the table (~$0.0022). August 20: **DeepSeek V4 Flash 0731** and **DeepSeek V4 Pro 0813** refreshed (footnote ¹⁹): current builds, AA Intelligence Index **51.8** / **53.2**, and peak/off-peak API pricing effective August 16. Pro's official GA agent benchmark update is documented in [docs/models/deepseek.md](models/deepseek.md). August 19: **GLM 5.3** added (footnote ¹⁸): Z.ai's new flagship (August 18, 2026), AA Intelligence Index **59.5**, priced identically to GLM 5.2 ($1.40 / $0.26 / $4.40). August 17: **GLM 5.2** OpenRouter promo refreshed (footnote ⁴): NovitaAI cut to 48% off; Open Sail Research now cheapest at $0.50 / $3.15. August 15: **Claude Fable 5** and **Qwen 3.6 Plus** added (footnotes ¹⁶–¹⁷); **Grok 4.5 / 4.6** and **Claude Fable 5** are now **GitHub Copilot native** (footnotes ¹², ¹⁴, ¹⁶). **Grok 4.6** and **Gemini 3.7 Flash** added August 14 (footnotes ¹⁴–¹⁵). **Gemini 3.6 Flash promotional price cut** to $0.75 / $3.75 (footnote ⁹). **Qwen 3.7 Max / 3.7 Plus price cuts** (see footnote ¹³). OpenRouter is running limited-time promos on GPT-5.6 Terra/Luna (⁸), GLM 5.2 (⁴), GLM 5.1, MiMo V2.5/V2.5 Pro, and MiniMax M3.
|
|
4
4
|
|
|
5
5
|
> **⏰ June 1, 2026 — GitHub Copilot switched to usage-based billing (AI Credits) today.**
|
|
6
6
|
>
|
|
@@ -24,6 +24,7 @@ All models are listed together below, sorted by Cost per intelligence ascending
|
|
|
24
24
|
|
|
25
25
|
| Model | Provider | Cost per intelligence | Intelligence Score | Input (per 1M) | Cached input | Output (per 1M) | Est. session | Context window |
|
|
26
26
|
| ----------------------------- | --------- | --------------------- | ------------------ | ----------------------------- | ----------------------------- | ----------------------------- | ------------ | -------------- |
|
|
27
|
+
| **GLM 5.3 Flash** ²⁰ | Z.ai | **~$0.0022** | **57.5** | $0.15 | $0.03 | $0.50 | ~$0.13 | 1M |
|
|
27
28
|
| **MiMo V2.5** | Xiaomi | **~$0.0026** | **38.0** | $0.14 | $0.0028 | $0.28 | ~$0.10 | 1M |
|
|
28
29
|
| **GPT-5.6 Luna** | OpenAI | **~$0.0042** | **52.3** | $0.20 | $0.02 | $1.20 | ~$0.22 | 1M |
|
|
29
30
|
| **MiniMax M3** | MiniMax | **~$0.0060** | **45.4** | $0.60 (≤512K) / $1.20 (>512K) | $0.12 (≤512K) / $0.24 (>512K) | $2.40 (≤512K) / $4.80 (>512K) | ~$0.27 | 1M |
|
|
@@ -31,6 +32,7 @@ All models are listed together below, sorted by Cost per intelligence ascending
|
|
|
31
32
|
| **MiMo V2.5 Pro** | Xiaomi | **~$0.0070** | **42.9** | $0.435 | $0.0036 | $0.87 | ~$0.30 | 1M |
|
|
32
33
|
| **Qwen 3.7 Plus** ¹³ | DashScope | **~$0.0074** | **39.4** | $0.32 (≤256K) / $0.40 (>256K) | $0.064 | $1.28 (≤256K) / $1.60 (>256K) | ~$0.29 | 1M |
|
|
33
34
|
| **MiniMax M3 Priority**⁵ | MiniMax | **~$0.0090** | **45.4** | $0.90 (≤512K) / $1.80 (>512K) | $0.18 (≤512K) / $0.36 (>512K) | $3.60 (≤512K) / $7.20 (>512K) | ~$0.41 | 1M |
|
|
35
|
+
| **Gemini 3.8 Flash** ²¹ | Google | **~$0.013** | **58.7** | $0.75 | $0.075 | $3.75 | ~$0.75 | 1M |
|
|
34
36
|
| **Gemini 3.7 Flash** ¹⁵ | Google | **~$0.013** | **56** | $0.75 | $0.075 | $3.75 | ~$0.75 | 1M |
|
|
35
37
|
| **Qwen 3.6 Plus** ¹⁷ | DashScope | **~$0.014** | **40.0** | $0.50 | $0.05 | $3.00 | ~$0.55 | 1M |
|
|
36
38
|
| **Gemini 3.6 Flash** ⁹ | Google | **~$0.015** | **51.6** | $0.75 | $0.075 | $3.75 | ~$0.75 | 1M |
|
|
@@ -58,7 +60,9 @@ All models are listed together below, sorted by Cost per intelligence ascending
|
|
|
58
60
|
| **GPT-5.6 Sol** | OpenAI | **~$0.090** | **60.9** | $5.00 | $0.50 | $30.00 | ~$5.50 | 1M |
|
|
59
61
|
| **Claude Opus 4.7** | Anthropic | **~$0.091** | **55.0** | $5.00 | $0.50 | $25.00 | ~$5.00 | 1M |
|
|
60
62
|
| **GPT-5.5** | OpenAI | **~$0.098** | **56.3** | $5.00 | $0.50 | $30.00 | ~$5.50 | 1M |
|
|
63
|
+
| **Claude Fable 5.1** ²³ | Anthropic | **~$0.15** | **65.7** | $10.00 | $0.25 | $50.00 | ~$10.00 | 1M |
|
|
61
64
|
| **Claude Fable 5** ¹⁶ | Anthropic | **~$0.161** | **62.1** | $10.00 | $1.00 | $50.00 | ~$10.00 | 1M |
|
|
65
|
+
| **Gemini 3.8 Flash Cyber** ²² | Google | — | — | — | — | — | — | — |
|
|
62
66
|
|
|
63
67
|
¹ Gemini 3.1 Pro pricing applies to prompts ≤200K tokens.
|
|
64
68
|
|
|
@@ -88,7 +92,7 @@ All models are listed together below, sorted by Cost per intelligence ascending
|
|
|
88
92
|
|
|
89
93
|
¹⁵ **Gemini 3.7 Flash** (Google, released **August 13, 2026**) — GitHub Copilot native (GA, Versatile) at the same promotional rates as 3.6 Flash. AA Intelligence Index **56** (high reasoning preset, #17/188). **Promotional pricing** of $0.75 / $0.075 / $3.75 per 1M input/cached/output through **December 31, 2026**, then $1.50 / $0.15 / $7.50 from January 1, 2027. 1M context window, text + image + speech + video input, and the fastest model on AA's leaderboard (340 t/s). See the [AA model page](https://artificialanalysis.ai/models/gemini-3-7-flash), [Google's Gemini page](https://deepmind.google/models/gemini/), and [GitHub Copilot pricing](https://docs.github.com/en/copilot/reference/copilot-billing/models-and-pricing).
|
|
90
94
|
|
|
91
|
-
¹⁶ **Claude Fable 5** (Anthropic, released **June 9, 2026**) is now a **GitHub Copilot native** model (GA, Powerful; Anthropic's new flagship tier, above Opus). AA Intelligence Index **62.1** (**#3/188**; AA page lists a rounded **62** on its [model page](https://artificialanalysis.ai/models/claude-fable-5)). Currently **#1 in Text, Agent, Code, and Overall Arena**. 1M context window, text + image input, adaptive reasoning. Priced at **$10.00 / $1.00 / $50.00** per 1M input/cached/output (90% cache discount; cache write $12.50). Uses the newer Claude tokenizer (~30% more tokens than pre-4.7 models). See the [GitHub Copilot models & pricing page](https://docs.github.com/en/copilot/reference/copilot-billing/models-and-pricing).
|
|
95
|
+
¹⁶ **Claude Fable 5** (Anthropic, released **June 9, 2026**) is now a **GitHub Copilot native** model (GA, Powerful; Anthropic's new flagship tier, above Opus). AA Intelligence Index **62.1** (**#3/188**; AA page lists a rounded **62** on its [model page](https://artificialanalysis.ai/models/claude-fable-5)). Currently **#1 in Text, Agent, Code, and Overall Arena**. 1M context window, text + image input, adaptive reasoning. Priced at **$10.00 / $1.00 / $50.00** per 1M input/cached/output (90% cache discount; cache write $12.50). Uses the newer Claude tokenizer (~30% more tokens than pre-4.7 models). **Superseded as Anthropic's flagship by Claude Fable 5.1 (September 1, 2026 — see footnote ²³)**. See the [GitHub Copilot models & pricing page](https://docs.github.com/en/copilot/reference/copilot-billing/models-and-pricing).
|
|
92
96
|
|
|
93
97
|
¹⁷ **Qwen 3.6 Plus** (DashScope, released **April 2, 2026**) — AA Intelligence Index **40.0**. Priced at **$0.50 / $0.05 / $3.00** per 1M input/cached/output (≤256K input tier; explicit cache read $0.05). 1M context window, text + image + video input, reasoning model. **Deprecated** — superseded by Qwen 3.7 Plus; AA marks it deprecated and no longer refreshes performance benchmarks beyond the default workload. See the [AA model page](https://artificialanalysis.ai/models/qwen3-6-plus).
|
|
94
98
|
|
|
@@ -96,13 +100,21 @@ All models are listed together below, sorted by Cost per intelligence ascending
|
|
|
96
100
|
|
|
97
101
|
¹⁹ **DeepSeek V4 Flash 0731** and **DeepSeek V4 Pro 0813** are the current API builds. Flash's AA Intelligence Index is **51.8** and Pro's is **53.2** (current OpenRouter/Artificial Analysis snapshots). Both are text-only with 1M context; Flash has 284B total / 13B active parameters and Pro has 1.6T total / 49B active. DeepSeek's official V4 pricing changed on August 16, 2026 to peak/off-peak rates: Flash is **$0.44 / $0.014 / $1.32** and Pro is **$1.32 / $0.044 / $3.96** per 1M input/cache-hit/output at peak, with off-peak rates at half those values. See the [DeepSeek pricing page](https://api-docs.deepseek.com/quick_start/pricing), [update log](https://api-docs.deepseek.com/updates/), and [canonical model record](models/deepseek.md).
|
|
98
102
|
|
|
103
|
+
²⁰ **GLM 5.3 Flash** (`glm-5.3-flash`) is Z.ai's first native multimodal model in the GLM-5 series (released **August 26, 2026**; tested pre-release as "ox-alpha"). Hybrid sparse + linear attention architecture with 320B total / 18B active parameters (open weights, MIT license) and a 1M-token context window; text + image input, text output. AA Intelligence Index **57.5** (OpenRouter's AA-sourced benchmark; AA's [model page](https://artificialanalysis.ai/models/glm-5-3-flash) lists a rounded **57** and ranks it **#3/110** within its open-weights size class), Coding **71.5**, Agentic **58.2**. Launch pricing is **$0.15** input / **$0.03** cached / **$0.50** output per 1M (list), with a **50% off** launch promo (effective **$0.075 / $0.015 / $0.25**) through **September 9, 2026** (24:00 UTC+8). At list price the modeled session cost is ~$0.13 and the cost per intelligence is ~$0.0022 — the cheapest row in this table. Always-thinking: `thinking.type` only supports `enabled`, `reasoning_effort` accepts `low` / `high` / `max` (default `max`). See the [Z.ai model page](https://docs.z.ai/guides/llm/glm-5.3-flash) and [AA model page](https://artificialanalysis.ai/models/glm-5-3-flash).
|
|
104
|
+
|
|
105
|
+
²¹ **Gemini 3.8 Flash** (`gemini-3.8-flash`) was released on **September 2, 2026** — Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and enterprise workflows. AA Intelligence Index **58.7** (OpenRouter's AA-sourced benchmark; AA's [model page](https://artificialanalysis.ai/models/gemini-3-8-flash) rounds to **59**, **#17/196**), Coding **76.3**, Agentic **50.0**. 1M context window (1,048,576 input / 65,536 output), text + image + speech + video + PDF input, text output; configurable thinking (`low`/`medium`/`high` — `minimal` is not supported). Fast at **305.5 t/s** per AA but **very verbose** (120M output tokens on the Intelligence Index vs a 71M median) with a high TTFT (13.39 s), so expect output-token costs above what the score alone suggests. **Not yet GitHub Copilot native** (GitHub's Google lineup still ends at 3.7 Flash) — priced from the Gemini API at the **same promotional rates as 3.7 Flash**: $0.75 / $0.075 / $3.75 per 1M input/cached/output through **December 31, 2026**, then $1.50 / $0.15 / $7.50 from January 1, 2027 (90% cache discount; Batch API 50% off; cache storage $0.50/1M per hour during the promo). Vendor-reported: HLE-Verified **54.9**. See the [Google announcement](https://blog.google/innovation-and-ai/models-and-research/gemini-models/3-8-flash-and-3-8-flash-cyber/), the [Gemini API model page](https://ai.google.dev/gemini-api/docs/models/gemini-3.8-flash), and the [AA model page](https://artificialanalysis.ai/models/gemini-3-8-flash).
|
|
106
|
+
|
|
107
|
+
²² **Gemini 3.8 Flash Cyber** was released on **September 2, 2026** as a cybersecurity-specialized variant sharing 3.8 Flash's foundational intelligence, focused on autonomous vulnerability discovery and automated patching (defensive capabilities prioritized over offensive ones). It is **not generally available**: access is limited to trusted defenders — government authorities, critical-infrastructure operators, and software maintainers — via Google's [Fairwind Program](https://deepmind.google/fairwind-program/). It has **no public API pricing** and **no Artificial Analysis Intelligence Index run**, so its cells show `—` (listed at the bottom per the table's sort rules) and it is not usable from VS Code custom endpoints. Vendor-reported results: CyberGym Pass@1 **86.2** (ahead of GPT-5.5-Cyber 85.6, Mythos 5 83.8, GPT-5.6 Sol 83.6, and Gemini 3.5 Flash Cyber 77.5), >70% success on Google's internal 20-language vulnerability-discovery benchmark (71.0 vs Gemini 3.7 Flash 58.9 and 3.5 Flash Cyber 46.6), CWE-Bench Pass@1 **47.2** (on the Pareto frontier vs Fable 5's 47.8 at significantly lower cost per rollout), and a **6.0** attack-success rate on the Gray Swan IPI prompt-injection benchmark (lower is better). Real-world reports: 2.6× more correct Chrome patches than larger commercial models; +7.5–9.7% pentest recall at 2.3–5.2× lower cost per Wiz. These are vendor-reported numbers, not independent replications. See the [Google announcement](https://blog.google/innovation-and-ai/models-and-research/gemini-models/3-8-flash-and-3-8-flash-cyber/) and the [model page](https://deepmind.google/models/gemini/cyber/).
|
|
108
|
+
|
|
109
|
+
²³ **Claude Fable 5.1** was released on **September 1, 2026** — Anthropic's new flagship, improving on Fable 5 across the board with the biggest gains in agentic coding, long-running agentic workflows, and knowledge work (long code refactors, front-end/visual code generation, finance and analysis), and tending to be more concise in plans and summaries. AA Intelligence Index **65.7** (OpenRouter's AA-sourced benchmark; AA's [model page](https://artificialanalysis.ai/models/claude-fable-5-1) rounds to **66** and ranks it **#1/196**) — the **highest score in this table** — with Coding **81.6** and Agentic **61.3**. GitHub Copilot native (GA, Powerful). 1M context window, text + image input, text output, adaptive reasoning; uses the newer Claude tokenizer (~30% more tokens than pre-4.7 models). Priced at **$10.00 / $0.25 / $50.00** per 1M input/cached/output — the same rates as Fable 5 except **cached input drops from $1.00 to $0.25** (97.5% cache discount); cache write $12.50. Modeled session cost stays ~$10.00, and CPI improves only to ~$0.15 (vs Fable 5's ~$0.161) — still the second-most expensive scored row. At max effort it is slow (66.2 t/s per AA) and very verbose (140M output tokens on the Intelligence Index vs a 71M median; TTFT 283.83 s). Like Fable 5, Anthropic retains request data by default, with zero-data-retention available through the end of 2026 under a time-bound exemption. See the [GitHub Copilot models & pricing page](https://docs.github.com/en/copilot/reference/copilot-billing/models-and-pricing) and the [AA model page](https://artificialanalysis.ai/models/claude-fable-5-1).
|
|
110
|
+
|
|
99
111
|
Cost per intelligence = estimated session cost ÷ Intelligence Index score. Session cost assumes ~10K input + ~2K output tokens per turn, 50 turns.
|
|
100
112
|
|
|
101
113
|
> **Notes:**
|
|
102
114
|
>
|
|
103
115
|
> - **DeepSeek V4** input pricing shown is the **peak cache miss** price. Peak cache hits are $0.014/M for Flash and $0.044/M for Pro; off-peak rates are half. See footnote ¹⁹ for the full schedule.
|
|
104
116
|
> - **MiMo** input pricing shown is the **cache miss** price. Cache hits are essentially free for V2.5 Pro ($0.0036/M, ~120× cheaper) and V2.5 ($0.0028/M, ~50× cheaper). A Xiaomi price cut took effect on 2026-05-27.
|
|
105
|
-
> - **Anthropic (Claude)** models also have a cache write cost ($6.25/MTok for Opus, $3.75/MTok for Sonnet, $1.25/MTok for Haiku). Opus 4.7+, Opus 5, Sonnet 5, and Fable 5 use a new tokenizer that produces approximately 30% more tokens for the same text.
|
|
117
|
+
> - **Anthropic (Claude)** models also have a cache write cost ($6.25/MTok for Opus, $3.75/MTok for Sonnet, $1.25/MTok for Haiku). Opus 4.7+, Opus 5, Sonnet 5, and Fable 5 / 5.1 use a new tokenizer that produces approximately 30% more tokens for the same text.
|
|
106
118
|
> - **OpenAI** models support cached input at 0.1× base input rate.
|
|
107
119
|
> - **Qwen** models use **tiered pricing** — determined by total input tokens per request. Prices above are for non-thinking mode.
|
|
108
120
|
> - **Kimi** official tables list **Cache Hit before Cache Miss** (opposite order to our table). The rows below transpose them so "Input" = cache miss and "Cached input" = cache hit.
|
|
@@ -110,13 +122,13 @@ Cost per intelligence = estimated session cost ÷ Intelligence Index score. Sess
|
|
|
110
122
|
> - **DashScope** offers a **free quota** of 1M input + 1M output tokens per model, valid for 90 days.
|
|
111
123
|
> - **MiniMax M3** uses **tiered pricing** — input price doubles above 512K input tokens. Cache hits are priced at 20% of the input rate ($0.12/M ≤512K, $0.24/M >512K). A **permanent 50% off** discount applies to all MiniMax-M3 pay-as-you-go usage (Standard and Priority tiers), making the effective rates half the list prices above.
|
|
112
124
|
> - **MiniMax M3 Priority** is not a separate model — it is the same `MiniMax-M3` weights invoked with `"service_tier": "priority"` in the request body. Priority costs **1.5× Standard** across input, output, and cache reads, in exchange for **priority admission** (faster responses, fewer failures during MiniMax peak hours — typically 15:00–17:30 weekdays). Capabilities, context window (1M, guaranteed 512K), vision, tool calling, rate limits (200 RPM / 10M TPM), and thinking modes are identical to Standard. See [docs/research/minimax-m3-priority.md](research/minimax-m3-priority.md) for the full breakdown.
|
|
113
|
-
> - **GLM** models support prompt caching — cache hits are priced at $0.26/M for 5.3, 5.2, and 5.1; $0.24/M for 5V Turbo.
|
|
125
|
+
> - **GLM** models support prompt caching — cache hits are priced at $0.26/M for 5.3, 5.2, and 5.1; $0.24/M for 5V Turbo; $0.03/M for 5.3 Flash (list; $0.015/M during the launch promo).
|
|
114
126
|
> - **MiMo** offers a **Token Plan** subscription model with discounted rates and a free cache-writing promotion.
|
|
115
127
|
> - For typical Copilot chat usage (short-to-medium prompts), you'll almost always fall in the lowest pricing tier.
|
|
116
128
|
|
|
117
129
|
> **How long does 7,000 credits last?** A Pro+ subscriber running 50-turn sessions could afford roughly **13 GPT-5.5 sessions** or **23 Opus sessions** per month — or mix and match. (Multiply session cost by 100 to convert to AI credits.)
|
|
118
130
|
|
|
119
|
-
> Prices last verified:
|
|
131
|
+
> Prices last verified: September 3, 2026. Always check the official pages for the latest rates:
|
|
120
132
|
>
|
|
121
133
|
> - [GitHub Copilot models & pricing](https://docs.github.com/en/copilot/reference/copilot-billing/models-and-pricing)
|
|
122
134
|
> - [OpenAI pricing](https://openai.com/api/pricing/)
|