copilot-custom-endpoint 1.9.3 → 2.1.0

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 CHANGED
@@ -27,6 +27,7 @@ That's it. No code, no servers to manage (unless the model specifically needs th
27
27
  | **MiMo V2.5 Pro** | Xiaomi | Optional (recommended) | ❌ | [Extension](https://marketplace.visualstudio.com/items?itemName=sdmapvstool.xiaomimimo-for-copilot) ★ / [Manual setup](docs/models/mimo.md) |
28
28
  | **Qwen 3.7 Plus** | DashScope | Optional (recommended) | ✅ | [Manual setup](docs/models/qwen.md) |
29
29
  | **Qwen 3.7 Max** | DashScope | Optional (recommended) | ❌ | [Manual setup](docs/models/qwen.md) |
30
+ | **Qwen 3.8 Max** | DashScope | Optional (recommended) | ✅ | [Manual setup](docs/models/qwen.md) |
30
31
  | **MiniMax M3** | MiniMax | No | ✅ | [Extension](https://github.com/tugudush/minimax-copilot) ★ / [Manual setup](docs/models/minimax.md) |
31
32
  | **GLM 5.2 / 5.1** | Z.ai | No | ✅ via extension proxy | [Manual setup](docs/models/glm.md) |
32
33
  | **GLM 5V Turbo** | Z.ai | No | ✅ | [Manual setup](docs/models/glm.md) |
@@ -110,6 +111,7 @@ All prices are **USD per 1M tokens** (non-cached). 1 AI credit = $0.01. To conve
110
111
  | ------------------------ | --------- | --------------------- | ------------------ | ------------ | ------ | -------------- |
111
112
  | **DeepSeek V4 Flash** | DeepSeek | **~$0.0025** | **40.3** | ~$0.10 | ✅ | 1M |
112
113
  | **MiMo V2.5** | Xiaomi | **~$0.0025** | **40** ³ | ~$0.10 | ✅ | 1M |
114
+ | **GPT-5.6 Luna** | OpenAI | **~$0.0043** | **51.2** | ~$0.22 | ✅ | 1M |
113
115
  | **MiniMax M3** | MiniMax | **~$0.0061** | **44.4** | ~$0.27 | ✅ | 1M |
114
116
  | **DeepSeek V4 Pro** | DeepSeek | **~$0.0068** | **44.3** | ~$0.30 | ✅ | 1M |
115
117
  | **MiMo V2.5 Pro** | Xiaomi | **~$0.0072** | **42.2** | ~$0.30 | ❌ | 1M |
@@ -118,59 +120,59 @@ All prices are **USD per 1M tokens** (non-cached). 1 AI credit = $0.01. To conve
118
120
  | **Kimi K2.6** | Moonshot | **~$0.021** | **42.8** | ~$0.88 | ✅ | 262K |
119
121
  | **Kimi K2.7 Code** | Moonshot | **~$0.021** | **42.0** | ~$0.88 | ✅ | 262K |
120
122
  | **GPT-5.4 mini** | OpenAI | **~$0.021** | **40.0** | ~$0.83 | ❌ | 400K |
121
- | **GPT-5.6 Luna** | OpenAI | **~$0.021** | **51.2** | ~$1.10 | ✅ | 1M |
122
123
  | **GLM 5.2** | Z.ai | **~$0.022** | **51.0** | ~$1.14 | ❌ | 1M |
123
124
  | **GLM 5.1** | Z.ai | **~$0.028** | **40.2** | ~$1.14 | ❌ | 200K |
124
125
  | **GLM 5V Turbo** | Z.ai | **~$0.029** | **34.0** ³ | ~$1.00 | ✅ | 200K |
125
126
  | **Gemini 3.6 Flash** ⁸ | Google | **~$0.030** | **50.0** | ~$1.50 | ✅ | 1M |
127
+ | **Qwen 3.8 Max** ¹⁰ | DashScope | **~$0.030** | **53.4** | ~$1.60 | ✅ | 1M |
126
128
  | **Gemini 3.5 Flash** | Google | **~$0.033** | **50.2** | ~$1.65 | ✅ | 1M |
129
+ | **GPT-5.6 Terra** | OpenAI | **~$0.040** | **55.0** | ~$2.20 | ✅ | 1M |
127
130
  | **Qwen 3.7 Max** | DashScope | **~$0.043** | **46.0** | ~$2.00 | ❌ | 1M |
128
131
  | **Gemini 3.1 Pro** | Google | **~$0.047** | **46.5** | ~$2.20 | ✅ | 1M |
129
- | **GPT-5.6 Terra** | OpenAI | **~$0.050** | **55.0** | ~$2.75 | ✅ | 1M |
130
132
  | **Kimi K3** ⁶ | Moonshot | **~$0.053** | **57.0** | ~$3.00 | ✅ | 1M |
131
133
  | **GPT-5.4** | OpenAI | **~$0.054** | **51.4** | ~$2.75 | ✅ | 1M |
132
134
  | **Claude Sonnet 5** ⁵ | Anthropic | **~$0.057** | **53.0** | ~$3.00 | ✅ | 1M |
133
135
  | **Claude Sonnet 4.6** | Anthropic | **~$0.064** | **47.2** | ~$3.00 | ✅ | 1M |
136
+ | **Claude Opus 5** ⁹ | Anthropic | **~$0.082** | **61.0** | ~$5.00 | ✅ | 1M |
134
137
  | **Claude Opus 4.8** | Anthropic | **~$0.090** | **55.7** | ~$5.00 | ✅ | 1M |
135
138
  | **Claude Opus 4.7** | Anthropic | **~$0.093** | **53.5** | ~$5.00 | ✅ | 1M |
136
139
  | **GPT-5.6 Sol** | OpenAI | **~$0.093** | **58.9** | ~$5.50 | ✅ | 1M |
137
140
  | **GPT-5.5** | OpenAI | **~$0.10** | **54.8** | ~$5.50 | ✅ | 1M |
138
- | **Raptor mini** | GitHub | — | — | ~$0.33 | ❌ | 264K |
139
- | **MAI-Code-1-Flash** ² | Microsoft | — | — | ~$0.83 | ❌ | — |
140
- | **Claude Haiku 4.5** | Anthropic | — | — | ~$1.00 | ❌ | 160K |
141
- | **GPT-5.3-Codex** | OpenAI | — | — | ~$2.28 | ❌ | 400K |
142
141
 
143
142
  ¹ Gemini 3.1 Pro pricing applies to prompts ≤200K tokens.
144
143
 
145
- ² MAI-Code-1-Flash is a continuously improving model — performance and behavior may evolve over time as new checkpoints are released.
146
-
147
144
  ³ Score is an **estimate** from Artificial Analysis (labelled "independent evaluation forthcoming"). Not a confirmed run of the full evaluation suite.
148
145
 
149
146
  ⁴ **MiniMax M3 Priority** is the same `MiniMax-M3` weights invoked with `"service_tier": "priority"` in the request body — **not a separate model**. Costs **1.5× Standard** (effective post-50%-off rates: $0.45 / $1.80 / $0.09 per 1M for input/output/cached ≤512K) in exchange for **priority admission** (faster responses, fewer failures during MiniMax peak hours). Capabilities, context window, vision, tools, rate limits, and thinking modes are identical to Standard. See [docs/research/minimax-m3-priority.md](docs/research/minimax-m3-priority.md).
150
147
 
151
148
  ⁵ **Claude Sonnet 5** has introductory pricing of **$2.00 / $10.00** per MTok (input/output) through **August 31, 2026**. Standard pricing of $3.00 / $15.00 shown above takes effect September 1, 2026. Released June 30, 2026 — Arena rankings pending. AA Intelligence Index score of **53.0** confirmed by [Artificial Analysis](https://artificialanalysis.ai/models/claude-sonnet-5).
152
149
 
153
- ⁶ **GPT-5.6** launched July 9, 2026. Sol, Terra, and Luna cost $5/$30, $2.50/$15, and $1/$6 per 1M input/output tokens respectively; all support image input and 1M context. Benchmark details and cache-write pricing are maintained in [docs/pricing.md](docs/pricing.md).
150
+ ⁶ **GPT-5.6** launched July 9, 2026. On July 30, OpenAI reduced Luna pricing by 80% to $0.20/$1.20 and Terra pricing by 20% to $2/$12 per 1M input/output tokens; Sol remains $5/$30. All support image input and 1M context. The Batch API offers an additional 50% discount for asynchronous jobs. Benchmark details and cache-write pricing are maintained in [docs/pricing.md](docs/pricing.md).
154
151
 
155
152
  ⁷ **Kimi K3** launched July 16, 2026. 2.8T params (open-source weights by July 27, 2026). Always-thinking reasoning model — uses `reasoning_effort` (not the K2.x `thinking` parameter). AA Intelligence Index score of **57.0** confirmed by [Artificial Analysis](https://artificialanalysis.ai/models/kimi-k3). Priced at $3.00 / $15.00 per MTok input/output. Requires the local Kimi proxy. See [docs/models/kimi.md](docs/models/kimi.md).
156
153
 
157
154
  ⁸ **Gemini 3.6 Flash** launched July 21, 2026. Priced at $1.50 / $7.50 per MTok input/output ($0.15 cached). AA Intelligence Index score of **50.0** confirmed by [Artificial Analysis](https://artificialanalysis.ai/models/gemini-3-6-flash). 1M context with vision.
158
155
 
156
+ ⁹ **Claude Opus 5** launched July 24, 2026. AA Intelligence Index score of **61.0** (#1 overall) confirmed by [Artificial Analysis](https://artificialanalysis.ai/models/claude-opus-5). Priced at $5.00 / $25.00 per MTok input/output (same as Opus 4.8). 1M context, text + image input, adaptive reasoning. Also available in Fast mode ($10/$50 per MTok input/output). Uses the newer Claude tokenizer (~30% more tokens than pre-4.7 models). See [Anthropic's Opus 5 announcement](https://www.anthropic.com/news/claude-opus-5).
157
+
158
+ ¹⁰ **Qwen 3.8 Max** (`qwen3.8-max`) has an OpenRouter-reported Artificial Analysis Intelligence Index of **53.4** (Coding **68.9**, Agentic **44.9**). Arena snapshots place it **#5 in Text**, **#2 in Vision**, and **#4 in Code/WebDev**; it is not listed in Agent Arena. See [OpenRouter](https://openrouter.ai/qwen/qwen3.8-max) and [Arena](https://arena.ai/leaderboard).
159
+
159
160
  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).
160
161
 
161
162
  > **👤 Personal picks** —
162
163
  >
163
- > For serious coding work, **MiniMax M3** is my top recommendation:
164
+ > For serious coding work, **GPT-5.6 Luna** is now my top value recommendation:
165
+ >
166
+ > - Higher intelligence score (**51.2**) than MiniMax M3 (44.4)
167
+ > - Lower modeled cost: ~$0.22/session versus MiniMax M3's ~$0.27
168
+ > - Available natively in Copilot, with reduced AI-credit consumption after OpenAI's July price cut
164
169
  >
165
- > - Highest intelligence score (44.4) among affordable models
166
- > - Only ~$0.27/session with vision + 1M context
167
- > - No local proxy required — works direct
168
- > **DeepSeek V4 Pro** is a strong backup — nearly identical intelligence (44.3 vs 44.4) and cost (~$0.0068 vs ~$0.0061 per intelligence point), but noticeably faster in practice.
170
+ > **MiniMax M3** remains the best direct non-OpenAI custom-endpoint option — no local proxy, vision, 1M context, and strong coding performance at ~$0.27/session.
169
171
  >
170
172
  > For **plan mode / architecture & design thinking**, **Kimi K3**, **GPT-5.6 Luna**, and **GLM 5.2** are worth the premium:
171
173
  >
172
174
  > - **Kimi K3** — new flagship. Highest intelligence score among custom-endpoint models (**57.0**, #3 overall). Best for complex reasoning, long-horizon planning, and architecture decisions. 1M context with vision. ~$3.00/session is ~11× MiniMax M3, so **not a daily driver** — reserve it for hard problems.
173
- > - **GPT-5.6 Luna** (Copilot native) — scores **51.2** at ~$1.10/session ($1/$6 per 1M tokens), making it the cheapest model with a 50+ intelligence score. Ideal if you have native Copilot access and want 50+ intelligence on a budget.
175
+ > - **GPT-5.6 Luna** (Copilot native) — scores **51.2** 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.
174
176
  > - **GLM 5.2** — scores **51.0**, reasoning-heavy planning at a lower price. 1M lossless context, ideal for whole-codebase planning and large spec review. ~$1.14/session is ~4× MiniMax M3, so **not a daily driver for agent mode** — reserve it for planning, then hand implementation off to a cheaper model.
175
177
  >
176
178
  > Cheaper alternatives for simpler tasks:
@@ -16,13 +16,25 @@ Here's a complete, real-world `chatLanguageModels.json` that combines **the `cus
16
16
  "apiKey": "",
17
17
  "apiType": "chat-completions",
18
18
  "models": [
19
+ {
20
+ "id": "qwen3.8-max",
21
+ "name": "Qwen 3.8 Max (vision)",
22
+ "url": "http://127.0.0.1:3458/v1/chat/completions",
23
+ "toolCalling": true,
24
+ "vision": true,
25
+ "streaming": true,
26
+ "maxInputTokens": 991000,
27
+ "maxOutputTokens": 131072
28
+ },
19
29
  {
20
30
  "id": "qwen3.7-max",
21
31
  "name": "Qwen 3.7 Max (text)",
22
32
  "url": "http://127.0.0.1:3458/v1/chat/completions",
23
33
  "toolCalling": true,
24
34
  "vision": false,
25
- "streaming": true
35
+ "streaming": true,
36
+ "maxInputTokens": 1048576,
37
+ "maxOutputTokens": 131072
26
38
  },
27
39
  {
28
40
  "id": "qwen3.7-plus",
@@ -30,7 +42,9 @@ Here's a complete, real-world `chatLanguageModels.json` that combines **the `cus
30
42
  "url": "http://127.0.0.1:3458/v1/chat/completions",
31
43
  "toolCalling": true,
32
44
  "vision": true,
33
- "streaming": true
45
+ "streaming": true,
46
+ "maxInputTokens": 1048576,
47
+ "maxOutputTokens": 131072
34
48
  }
35
49
  ]
36
50
  },
@@ -127,7 +141,7 @@ Here's a complete, real-world `chatLanguageModels.json` that combines **the `cus
127
141
  "models": [
128
142
  {
129
143
  "id": "MiniMax-M3",
130
- "name": "MiniMax M3 (Standard)",
144
+ "name": "MiniMax M3 (vision)",
131
145
  "url": "https://api.minimax.io/v1/chat/completions",
132
146
  "toolCalling": true,
133
147
  "vision": true,
@@ -1,6 +1,6 @@
1
1
  # Qwen (DashScope) — VS Code Custom Endpoint Setup Guide
2
2
 
3
- > **TL;DR:** `qwen3.7-plus` (vision) and `qwen3.7-max` (text) work both direct and via the local proxy. The proxy gives you dynamic thinking suppression: reasoning stays ON in plain chat but turns OFF automatically when tools are invoked. The direct path is simpler if you don't need reasoning in chat.
3
+ > **TL;DR:** `qwen3.8-max` (vision), `qwen3.7-plus` (vision), and `qwen3.7-max` (text) work both direct and via the local proxy. The proxy gives you dynamic thinking suppression: reasoning stays ON in plain chat but turns OFF automatically when tools are invoked. The direct path is simpler if you don't need reasoning in chat.
4
4
 
5
5
  ## At a Glance
6
6
 
@@ -8,10 +8,10 @@
8
8
  | ---------------------- | -------------------------------------------------------------------------------- |
9
9
  | Mode | **Proxy** (local on `:3458`) **or** **Direct** (static `enable_thinking: false`) |
10
10
  | Billing | **Pay-as-You-Go only** — 1M-token free quota for new users |
11
- | Vision | ✅ Yes (`qwen3.7-plus`) |
11
+ | Vision | ✅ Yes (`qwen3.8-max`, `qwen3.7-plus`) |
12
12
  | Tool calling | ✅ Yes |
13
13
  | Context | 1M |
14
- | Max output | _see Models table_ |
14
+ | Max output | 131K (`qwen3.8-max`); see Models table for other models |
15
15
  | Endpoint | `https://dashscope-intl.aliyuncs.com/compatible-mode/v1/chat/completions` |
16
16
  | Proxy endpoint | `http://127.0.0.1:3458/v1/chat/completions` |
17
17
  | Auth | `Authorization: Bearer $DASHSCOPE_API_KEY` |
@@ -20,10 +20,11 @@
20
20
 
21
21
  ### Models
22
22
 
23
- | Model | Vision | Context | Max output | Notes |
24
- | -------------- | ------ | ------- | --------------- | -------------------------------------- |
25
- | `qwen3.7-plus` | ✅ Yes | 1M | _(unspecified)_ | Primary model with image understanding |
26
- | `qwen3.7-max` | ❌ No | 1M | _(unspecified)_ | Larger text-only model |
23
+ | Model | Vision | Context | Max output | Notes |
24
+ | -------------- | ------ | ------- | --------------- | ------------------------------------------------------- |
25
+ | `qwen3.8-max` | ✅ Yes | 1M | 131K | Frontier multimodal model; reasoning enabled by default |
26
+ | `qwen3.7-plus` | ✅ Yes | 1M | _(unspecified)_ | Primary model with image understanding |
27
+ | `qwen3.7-max` | ❌ No | 1M | _(unspecified)_ | Larger text-only model |
27
28
 
28
29
  > The live `chatLanguageModels.json` points Qwen at the local proxy by default; the direct DashScope URL is shown below for users who prefer a static `enable_thinking: false` setup.
29
30
 
@@ -42,7 +43,7 @@
42
43
  1. **Use the direct-path JSON snippet** below.
43
44
  2. **Set your `DASHSCOPE_API_KEY`** via Command Palette → **Chat: Manage Language Models**.
44
45
  3. **Configure the Utility Small Model** — Open Settings → search **"Chat: Utility Small Model"** → pick your fastest model (e.g., DeepSeek V4 Flash or MiMo V2.5). [Why?](../../README.md#4-configure-the-utility-small-model)
45
- 4. **Restart VS Code** and pick "Qwen 3.7 Plus" or "Qwen 3.7 Max".
46
+ 4. **Restart VS Code** and pick "Qwen 3.8 Max", "Qwen 3.7 Plus", or "Qwen 3.7 Max".
46
47
 
47
48
  ## Setup
48
49
 
@@ -73,6 +74,17 @@ DashScope is region-specific — your API key only works on the endpoint it was
73
74
  "apiKey": "",
74
75
  "apiType": "chat-completions",
75
76
  "models": [
77
+ {
78
+ "id": "qwen3.8-max",
79
+ "name": "Qwen 3.8 Max (vision)",
80
+ "url": "https://dashscope-intl.aliyuncs.com/compatible-mode/v1/chat/completions",
81
+ "toolCalling": true,
82
+ "vision": true,
83
+ "streaming": true,
84
+ "maxInputTokens": 991000,
85
+ "maxOutputTokens": 131072,
86
+ "requestBody": { "enable_thinking": false }
87
+ },
76
88
  {
77
89
  "id": "qwen3.7-max",
78
90
  "name": "Qwen 3.7 Max (text)",
@@ -120,6 +132,16 @@ DashScope is region-specific — your API key only works on the endpoint it was
120
132
  "apiKey": "",
121
133
  "apiType": "chat-completions",
122
134
  "models": [
135
+ {
136
+ "id": "qwen3.8-max",
137
+ "name": "Qwen 3.8 Max (vision)",
138
+ "url": "http://127.0.0.1:3458/v1/chat/completions",
139
+ "toolCalling": true,
140
+ "vision": true,
141
+ "streaming": true,
142
+ "maxInputTokens": 991000,
143
+ "maxOutputTokens": 131072
144
+ },
123
145
  {
124
146
  "id": "qwen3.7-max",
125
147
  "name": "Qwen 3.7 Max (text)",
@@ -195,7 +217,8 @@ The proxy detects active tool use by examining the conversation state (a `"tool"
195
217
 
196
218
  ## Notes
197
219
 
198
- - **Vision (`qwen3.7-plus`)** uses OpenAI-compatible `content` array format. Base64 data URIs work reliably; external image URLs may fail if DashScope can't reach them. If a drag-and-drop image fails to load, providing the absolute file path (e.g. `c:\path\to\image.png`) in the prompt is a reliable workaround.
220
+ - **Vision (`qwen3.8-max`, `qwen3.7-plus`)** uses OpenAI-compatible `content` array format. Base64 data URIs work reliably; external image URLs may fail if DashScope can't reach them. If a drag-and-drop image fails to load, providing the absolute file path (e.g. `c:\path\to\image.png`) in the prompt is a reliable workaround.
221
+ - **Qwen 3.8 reasoning:** `qwen3.8-max` enables reasoning by default and supports `reasoning_effort` values `low`, `medium`, and `xhigh` (default). The direct snippet disables reasoning for stable VS Code tool loops; the proxy leaves it on for plain chat and suppresses it when tool activity is detected.
199
222
  - **Thinking trade-off:** Direct = thinking always off (loops stable, no reasoning visible). Proxy = thinking on in plain chat, off in tool turns.
200
223
  - **`tool_choice` only supports `auto`** — don't override it (VS Code's default is `auto`).
201
224
 
@@ -216,6 +239,7 @@ For the cross-provider comparison, see [docs/pricing.md](../pricing.md). DashSco
216
239
 
217
240
  | Model | Input (≤ 256K tokens) | Input (> 256K tokens) | Output (≤ 256K tokens) | Output (> 256K tokens) |
218
241
  | -------------- | --------------------- | --------------------- | ---------------------- | ---------------------- |
242
+ | `qwen3.8-max` | $2.00 | — | $6.00 | — |
219
243
  | `qwen3.7-plus` | $0.40 / 1M | $1.20 / 1M | $1.60 / 1M | $4.80 / 1M |
220
244
  | `qwen3.7-max` | $2.50 / 1M (≤ 1M) | — | $7.50 / 1M (≤ 1M) | — |
221
245
 
package/docs/pricing.md CHANGED
@@ -1,6 +1,6 @@
1
1
  # Pricing
2
2
 
3
- > **Updated:** July 26, 2026 — Claude Opus 5 added (released July 24, 2026). AA Intelligence Index **61.0** (#1 overall), pricing $5.00 / $25.00 per MTok input/output.
3
+ > **Updated:** August 3, 2026 — GPT-5.6 Luna and Terra price reductions verified against OpenAI's July 30 announcement. Luna is 80% cheaper and Terra is 20% cheaper; Sol pricing is unchanged.
4
4
 
5
5
  > **⏰ June 1, 2026 — GitHub Copilot switched to usage-based billing (AI Credits) today.**
6
6
  >
@@ -26,6 +26,7 @@ All models are listed together below, sorted by Cost per intelligence ascending
26
26
  | ------------------------ | --------- | --------------------- | ------------------ | ----------------------------- | ----------------------------- | ----------------------------- | ------------ | -------------- |
27
27
  | **DeepSeek V4 Flash** | DeepSeek | **~$0.0025** | **40.3** | $0.14 | $0.0028 | $0.28 | ~$0.10 | 1M |
28
28
  | **MiMo V2.5** | Xiaomi | **~$0.0025** | **40** ³ | $0.14 | $0.0028 | $0.28 | ~$0.10 | 1M |
29
+ | **GPT-5.6 Luna** | OpenAI | **~$0.0043** | **51.2** | $0.20 | $0.02 | $1.20 | ~$0.22 | 1M |
29
30
  | **MiniMax M3** | MiniMax | **~$0.0061** | **44.4** | $0.60 (≤512K) / $1.20 (>512K) | $0.12 (≤512K) / $0.24 (>512K) | $2.40 (≤512K) / $4.80 (>512K) | ~$0.27 | 1M |
30
31
  | **DeepSeek V4 Pro** | DeepSeek | **~$0.0068** | **44.3** | $0.435 | $0.003625 | $0.87 | ~$0.30 | 1M |
31
32
  | **MiMo V2.5 Pro** | Xiaomi | **~$0.0072** | **42.2** | $0.435 | $0.0036 | $0.87 | ~$0.30 | 1M |
@@ -34,16 +35,15 @@ All models are listed together below, sorted by Cost per intelligence ascending
34
35
  | **Kimi K2.6** | Moonshot | **~$0.021** | **42.8** | $0.95 | $0.16 | $4.00 | ~$0.88 | 256K |
35
36
  | **Kimi K2.7 Code** | Moonshot | **~$0.021** | **42.0** | $0.95 | $0.19 | $4.00 | ~$0.88 | 262K |
36
37
  | **GPT-5.4 mini** | OpenAI | **~$0.021** | **40.0** | $0.75 | $0.075 | $4.50 | ~$0.83 | 400K |
37
- | **GPT-5.6 Luna** | OpenAI | **~$0.021** | **51.2** | $1.00 | $0.10 | $6.00 | ~$1.10 | 1M |
38
38
  | **GLM 5.2** ⁴ | Z.ai | **~$0.022** | **51.0** | $1.40 | $0.26 | $4.40 | ~$1.14 | 1M |
39
39
  | **GLM 5.1** | Z.ai | **~$0.028** | **40.2** | $1.40 | $0.26 | $4.40 | ~$1.14 | 200K |
40
40
  | **GLM 5V Turbo** | Z.ai | **~$0.029** | **34.0** ³ | $1.20 | $0.24 | $4.00 | ~$1.00 | 200K |
41
41
  | **Gemini 3.6 Flash** ⁹ | Google | **~$0.030** | **50.0** | $1.50 | $0.15 | $7.50 | ~$1.50 | 1M |
42
+ | **Qwen 3.8 Max** ¹¹ | DashScope | **~$0.030** | **53.4** | $2.00 | $0.25 | $6.00 | ~$1.60 | 1M |
42
43
  | **Gemini 3.5 Flash** | Google | **~$0.033** | **50.2** | $1.50 | $0.15 | $9.00 | ~$1.65 | 1M |
44
+ | **GPT-5.6 Terra** | OpenAI | **~$0.040** | **55.0** | $2.00 | $0.20 | $12.00 | ~$2.20 | 1M |
43
45
  | **Qwen 3.7 Max** | DashScope | **~$0.043** | **46.0** | $2.50 (≤1M) | — | $7.50 (≤1M) | ~$2.00 | 1M |
44
46
  | **Gemini 3.1 Pro** | Google | **~$0.047** | **46.5** | $2.00¹ | $0.20 | $12.00¹ | ~$2.20 | 1M |
45
- | **GPT-5.6 Terra** | OpenAI | **~$0.050** | **55.0** | $2.50 | $0.25 | $15.00 | ~$2.75 | 1M |
46
- | **GPT-5.3-Codex** | OpenAI | **~$0.052** | **44.0** ³ | $1.75 | $0.175 | $14.00 | ~$2.28 | 400K |
47
47
  | **Kimi K3** ⁷ | Moonshot | **~$0.053** | **57.0** | $3.00 | $0.30 | $15.00 | ~$3.00 | 1M |
48
48
  | **GPT-5.4** | OpenAI | **~$0.054** | **51.4** | $2.50 | $0.25 | $15.00 | ~$2.75 | 1M |
49
49
  | **Claude Sonnet 5** ⁶ | Anthropic | **~$0.057** | **53.0** | $3.00 | $0.30 | $15.00 | ~$3.00 | 1M |
@@ -53,14 +53,9 @@ All models are listed together below, sorted by Cost per intelligence ascending
53
53
  | **Claude Opus 4.7** | Anthropic | **~$0.093** | **53.5** | $5.00 | $0.50 | $25.00 | ~$5.00 | 1M |
54
54
  | **GPT-5.6 Sol** | OpenAI | **~$0.093** | **58.9** | $5.00 | $0.50 | $30.00 | ~$5.50 | 1M |
55
55
  | **GPT-5.5** | OpenAI | **~$0.10** | **54.8** | $5.00 | $0.50 | $30.00 | ~$5.50 | 1M |
56
- | **Raptor mini** | GitHub | — | — | $0.25 | $0.025 | $2.00 | ~$0.33 | 264K |
57
- | **MAI-Code-1-Flash** ² | Microsoft | — | — | $0.75 | $0.075 | $4.50 | ~$0.83 | — |
58
- | **Claude Haiku 4.5** | Anthropic | — | — | $1.00 | $0.10 | $5.00 | ~$1.00 | 160K |
59
56
 
60
57
  ¹ Gemini 3.1 Pro pricing applies to prompts ≤200K tokens.
61
58
 
62
- ² MAI-Code-1-Flash is a continuously improving model — performance and behavior may evolve over time as new checkpoints are released.
63
-
64
59
  ³ Score is an **estimate** from Artificial Analysis (labelled "independent evaluation forthcoming"). Not a confirmed run of the full evaluation suite.
65
60
 
66
61
  ⁴ GLM 5.2 is Z.ai's newly released flagship model (Opus-level intelligence / 1M Solid context). AA Intelligence Index score (**51.0**) and pricing of **$1.40** (input), **$0.26** (cached), and **$4.40** (output) per 1M tokens confirmed on its [model page](https://artificialanalysis.ai/models/glm-5-2). Session cost is identical to GLM 5.1 (~$1.14).
@@ -71,12 +66,14 @@ All models are listed together below, sorted by Cost per intelligence ascending
71
66
 
72
67
  ⁷ **Kimi K3** was released on **July 16, 2026**. 2.8T parameters (open-source, weights by July 27, 2026). 1M context window. Always-thinking reasoning model — uses `reasoning_effort` (not the K2.x `thinking` parameter). Fixed sampling: `temperature=1`, `top_p=0.95`. Pricing is flat (no tiering by context length). See [Kimi K3 pricing](https://platform.kimi.ai/docs/pricing/chat-k3) and the [Artificial Analysis model page](https://artificialanalysis.ai/models/kimi-k3).
73
68
 
74
- ⁸ **GPT-5.6** pricing is **$1.00 / $0.10 / $6.00** for Luna, **$2.50 / $0.25 / $15.00** for Terra, and **$5.00 / $0.50 / $30.00** for Sol (input / cached input / output per 1M tokens). OpenAI bills cache writes at 1.25x the uncached input rate for GPT-5.6 and later. All three tiers have a 1M-token context window and support text + image input. See [OpenAI's GPT-5.6 announcement](https://openai.com/index/gpt-5-6/) and the [Artificial Analysis model pages](https://artificialanalysis.ai/models/gpt-5-6-sol).
69
+ ⁸ **GPT-5.6** pricing is **$0.20 / $0.02 / $1.20** for Luna, **$2.00 / $0.20 / $12.00** for Terra, and **$5.00 / $0.50 / $30.00** for Sol (input / cached input / output per 1M tokens). OpenAI reduced Luna by 80% and Terra by 20% on July 30, 2026; Sol is unchanged. OpenAI bills cache writes at 1.25x the uncached input rate for GPT-5.6 and later. The Batch API provides an additional 50% discount for asynchronous jobs. All three tiers have a 1M-token context window and support text + image input. See [OpenAI's GPT-5.6 price update](https://openai.com/index/advancing-the-price-performance-frontier-with-gpt-5-6/) and [GPT-5.6 announcement](https://openai.com/index/gpt-5-6/).
75
70
 
76
71
  ⁹ **Gemini 3.6 Flash** was released on **July 21, 2026**. AA Intelligence Index score (**50.0**) confirmed by [Artificial Analysis](https://artificialanalysis.ai/models/gemini-3-6-flash). Pricing is $1.50 input / $0.15 cached / $7.50 output per MTok.
77
72
 
78
73
  ¹⁰ **Claude Opus 5** was released on **July 24, 2026**. AA Intelligence Index score (**61.0**) confirmed by [Artificial Analysis](https://artificialanalysis.ai/models/claude-opus-5). Pricing is $5.00 / $0.50 / $25.00 per MTok input/cached/output (same as Opus 4.8). 1M context window, text + image input, adaptive reasoning. Also available in Fast mode ($10/$50 per MTok input/output) and Batch ($2.50/$12.50 per MTok input/output). Uses the newer Claude tokenizer (~30% more tokens than pre-4.7 models). See [Anthropic's Opus 5 announcement](https://www.anthropic.com/news/claude-opus-5).
79
74
 
75
+ ¹¹ **Qwen 3.8 Max** (`qwen3.8-max`) launched on **August 3, 2026**. Qwen Cloud lists $2.00 input / $6.00 output per 1M tokens, $0.25 implicit-cache input, $2.50 explicit-cache creation, and $0.17 explicit-cache reads. The modeled session cost is ~$1.60 using 10K input + 2K output tokens per turn across 50 turns. It has a 1M context window, 131K maximum output, and text/image/video input. OpenRouter currently reports an Artificial Analysis Intelligence Index of **53.4** for this model. See the [Qwen Cloud model page](https://www.qwencloud.com/models/qwen3.8-max) and [OpenRouter model page](https://openrouter.ai/qwen/qwen3.8-max).
76
+
80
77
  Cost per intelligence = estimated session cost ÷ Intelligence Index score. Session cost assumes ~10K input + ~2K output tokens per turn, 50 turns.
81
78
 
82
79
  > **Notes:**
@@ -93,17 +90,16 @@ Cost per intelligence = estimated session cost ÷ Intelligence Index score. Sess
93
90
  > - **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.
94
91
  > - **GLM** models support prompt caching — cache hits are priced at $0.24/M for 5V Turbo and $0.26/M for 5.1.
95
92
  > - **MiMo** offers a **Token Plan** subscription model with discounted rates and a free cache-writing promotion.
96
- > - **MAI-Code-1-Flash** is a continuously improving model — performance and behavior may evolve over time as new checkpoints are released.
97
93
  > - For typical Copilot chat usage (short-to-medium prompts), you'll almost always fall in the lowest pricing tier.
98
94
 
99
- > **How long does 7,000 credits last?** A Pro+ subscriber running 50-turn sessions could afford roughly **13 GPT-5.5 sessions**, **23 Opus sessions**, or **212 Raptor mini sessions** per month — or mix and match. (Multiply session cost by 100 to convert to AI credits.)
95
+ > **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.)
100
96
 
101
- > Prices last verified: July 26, 2026. Always check the official pages for the latest rates:
97
+ > Prices last verified: August 3, 2026. Always check the official pages for the latest rates:
102
98
  >
103
99
  > - [GitHub Copilot models & pricing](https://docs.github.com/en/copilot/reference/copilot-billing/models-and-pricing)
104
- > - [Microsoft MAI-Code-1-Flash model card](https://docs.github.com/en/copilot/reference/ai-models/model-comparison#task-general-purpose-coding-and-writing)
105
100
  > - [OpenAI pricing](https://openai.com/api/pricing/)
106
101
  > - [OpenAI GPT-5.6 announcement](https://openai.com/index/gpt-5-6/)
102
+ > - [OpenAI GPT-5.6 price update](https://openai.com/index/advancing-the-price-performance-frontier-with-gpt-5-6/)
107
103
  > - [Anthropic (Claude) pricing](https://platform.claude.com/docs/en/about-claude/pricing)
108
104
  > - [Google Gemini pricing](https://ai.google.dev/pricing)
109
105
  > - [DashScope pricing](https://www.alibabacloud.com/help/en/model-studio/billing-for-model-studio)
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "copilot-custom-endpoint",
3
- "version": "1.9.3",
3
+ "version": "2.1.0",
4
4
  "description": "Local proxies for VS Code Copilot custom endpoints — Kimi K2, Qwen 3.x & MiMo V2.5",
5
5
  "license": "MIT",
6
6
  "type": "module",
@@ -57,4 +57,4 @@
57
57
  "dependencies": {
58
58
  "dotenv": "^17.4.2"
59
59
  }
60
- }
60
+ }
@@ -5,7 +5,7 @@ import { createProxy } from '../lib/create-proxy.mjs'
5
5
 
6
6
  /**
7
7
  * Supported model scope for this proxy:
8
- * - Validated with `qwen3.7-plus` and `qwen3.7-max`.
8
+ * - Validated with `qwen3.8-max`, `qwen3.7-plus`, and `qwen3.7-max`.
9
9
  * - Expected to work for any Qwen3 hybrid-thinking model (qwen3-* series)
10
10
  * that supports the `enable_thinking` top-level field on DashScope's
11
11
  * OpenAI-compatible surface.