pi-hypercharm-provider 1.1.1 → 1.1.3

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
@@ -26,19 +26,19 @@ _Hyperoptimized coding models — DeepSeek, GLM, Kimi, Qwen, MiniMax, Gemma, GPT
26
26
 
27
27
  | Model | Type | Context | Max Tokens | Input Cost | Output Cost |
28
28
  |-------|------|---------|------------|------------|-------------|
29
- | DeepSeek V4 Flash | Text | 1.0M | 384K | $0.14 | $0.28 |
30
- | DeepSeek V4 Pro | Text | 1.0M | 384K | $1.74 | $3.48 |
31
- | Gemma 4 26B A4B | Text + Image | 256K | 26K | $0.12 | $0.43 |
29
+ | DeepSeek V4 Flash | Text | 1.0M | 384K | $0.20 | $0.40 |
30
+ | DeepSeek V4 Pro | Text | 1.0M | 384K | $2.40 | $4.80 |
31
+ | Gemma 4 26B A4B | Text + Image | 256K | 26K | $0.12 | $0.41 |
32
32
  | GLM 5.1 | Text + Image | 203K | 131K | $1.40 | $4.40 |
33
- | GLM-5 | Text | 203K | 20K | $0.77 | $2.39 |
34
- | GLM-5.2 | Text | 1.0M | 131K | $1.40 | $4.40 |
35
- | GPT-OSS 120B | Text | 131K | 13K | $0.18 | $0.61 |
36
- | Kimi K2.5 | Text + Image | 262K | 26K | $0.48 | $2.54 |
33
+ | GLM-5 | Text | 203K | 20K | $0.81 | $2.59 |
34
+ | GLM-5.2 | Text | 1.0M | 128K | $1.40 | $4.40 |
35
+ | GPT-OSS 120B | Text | 131K | 13K | $0.17 | $0.56 |
36
+ | Kimi K2.5 | Text + Image | 262K | 26K | $0.52 | $2.72 |
37
37
  | Kimi K2.6 | Text + Image | 262K | 262K | $0.95 | $4.00 |
38
- | Kimi K2.7 Code | Text | 262K | 262K | $0.95 | $4.00 |
38
+ | Kimi K2.7 Code | Text | 256K | 16K | $0.95 | $4.00 |
39
39
  | Llama 3.3 70B Instruct | Text | 128K | 13K | $0.61 | $1.04 |
40
- | Llama 4 Maverick 17B 128E FP8 | Text | 430K | 43K | $0.27 | $0.89 |
41
- | MiniMax M2.7 | Text | 205K | 20K | $0.82 | $2.64 |
40
+ | Llama 4 Maverick 17B 128E FP8 | Text | 430K | 43K | $0.28 | $0.93 |
41
+ | MiniMax M2.7 | Text | 205K | 20K | $0.42 | $1.56 |
42
42
  | Qwen3 Coder 480B A35B INT4 | Text | 106K | 11K | $0.60 | $2.08 |
43
43
  | Qwen3 Next 80B A3B | Text | 262K | 26K | $0.12 | $1.14 |
44
44
  | Qwen3.6-Flash | Text + Image | 1.0M | 64K | $1.00 | $4.00 |
package/index.ts CHANGED
@@ -48,7 +48,7 @@ interface JsonModel {
48
48
  id: string;
49
49
  name: string;
50
50
  reasoning: boolean;
51
- input: string[];
51
+ input: ("text" | "image")[];
52
52
  cost: {
53
53
  input: number;
54
54
  output: number;
@@ -71,7 +71,7 @@ interface JsonModel {
71
71
  interface PatchEntry {
72
72
  name?: string;
73
73
  reasoning?: boolean;
74
- input?: string[];
74
+ input?: ("text" | "image")[];
75
75
  cost?: {
76
76
  input?: number;
77
77
  output?: number;
@@ -178,7 +178,7 @@ function buildModels(base: JsonModel[], custom: JsonModel[], patch: PatchData):
178
178
  }
179
179
  if (!model.thinkingLevelMap) {
180
180
  model.thinkingLevelMap = {
181
- minimal: null, low: null, medium: null, high: "high", xhigh: "max", max: "max",
181
+ minimal: null, low: null, medium: null, high: "high", max: "max",
182
182
  };
183
183
  }
184
184
  }
package/models.json CHANGED
@@ -7,10 +7,10 @@
7
7
  "text"
8
8
  ],
9
9
  "cost": {
10
- "input": 0.14,
11
- "output": 0.28,
10
+ "input": 0.2,
11
+ "output": 0.4,
12
12
  "cacheRead": 0.03,
13
- "cacheWrite": 0.03
13
+ "cacheWrite": 0.04
14
14
  },
15
15
  "contextWindow": 1000000,
16
16
  "maxTokens": 384000,
@@ -19,7 +19,6 @@
19
19
  "low": null,
20
20
  "medium": null,
21
21
  "high": "high",
22
- "xhigh": "max",
23
22
  "max": "max"
24
23
  },
25
24
  "compat": {
@@ -39,10 +38,10 @@
39
38
  "text"
40
39
  ],
41
40
  "cost": {
42
- "input": 1.74,
43
- "output": 3.48,
41
+ "input": 2.4,
42
+ "output": 4.8,
44
43
  "cacheRead": 0.15,
45
- "cacheWrite": 0.15
44
+ "cacheWrite": 0.2
46
45
  },
47
46
  "contextWindow": 1000000,
48
47
  "maxTokens": 384000,
@@ -51,7 +50,6 @@
51
50
  "low": null,
52
51
  "medium": null,
53
52
  "high": "high",
54
- "xhigh": "max",
55
53
  "max": "max"
56
54
  },
57
55
  "compat": {
@@ -72,9 +70,9 @@
72
70
  "image"
73
71
  ],
74
72
  "cost": {
75
- "input": 0.12,
76
- "output": 0.426,
77
- "cacheRead": 0.06,
73
+ "input": 0.116,
74
+ "output": 0.408,
75
+ "cacheRead": 0.058,
78
76
  "cacheWrite": 0
79
77
  },
80
78
  "contextWindow": 256000,
@@ -105,9 +103,9 @@
105
103
  "text"
106
104
  ],
107
105
  "cost": {
108
- "input": 0.77,
109
- "output": 2.388,
110
- "cacheRead": 0.385,
106
+ "input": 0.81,
107
+ "output": 2.592,
108
+ "cacheRead": 0.405,
111
109
  "cacheWrite": 0
112
110
  },
113
111
  "contextWindow": 202752,
@@ -124,10 +122,10 @@
124
122
  "input": 1.4,
125
123
  "output": 4.4,
126
124
  "cacheRead": 0.26,
127
- "cacheWrite": 0.14
125
+ "cacheWrite": 0.28
128
126
  },
129
- "contextWindow": 1048576,
130
- "maxTokens": 131072
127
+ "contextWindow": 1000000,
128
+ "maxTokens": 128000
131
129
  },
132
130
  {
133
131
  "id": "gpt-oss-120b",
@@ -137,9 +135,9 @@
137
135
  "text"
138
136
  ],
139
137
  "cost": {
140
- "input": 0.18,
141
- "output": 0.61,
142
- "cacheRead": 0.09,
138
+ "input": 0.168,
139
+ "output": 0.562,
140
+ "cacheRead": 0.084,
143
141
  "cacheWrite": 0
144
142
  },
145
143
  "contextWindow": 131072,
@@ -154,9 +152,9 @@
154
152
  "image"
155
153
  ],
156
154
  "cost": {
157
- "input": 0.477,
158
- "output": 2.535,
159
- "cacheRead": 0.2385,
155
+ "input": 0.518,
156
+ "output": 2.72,
157
+ "cacheRead": 0.259,
160
158
  "cacheWrite": 0
161
159
  },
162
160
  "contextWindow": 262144,
@@ -192,8 +190,8 @@
192
190
  "cacheRead": 0.16,
193
191
  "cacheWrite": 0.19
194
192
  },
195
- "contextWindow": 262000,
196
- "maxTokens": 262000
193
+ "contextWindow": 256000,
194
+ "maxTokens": 16000
197
195
  },
198
196
  {
199
197
  "id": "llama-3.3-70b-instruct",
@@ -219,9 +217,9 @@
219
217
  "text"
220
218
  ],
221
219
  "cost": {
222
- "input": 0.274,
223
- "output": 0.894,
224
- "cacheRead": 0.137,
220
+ "input": 0.284,
221
+ "output": 0.934,
222
+ "cacheRead": 0.142,
225
223
  "cacheWrite": 0
226
224
  },
227
225
  "contextWindow": 430000,
@@ -235,9 +233,9 @@
235
233
  "text"
236
234
  ],
237
235
  "cost": {
238
- "input": 0.82,
239
- "output": 2.64,
240
- "cacheRead": 0.41,
236
+ "input": 0.42,
237
+ "output": 1.56,
238
+ "cacheRead": 0.21,
241
239
  "cacheWrite": 0.06
242
240
  },
243
241
  "contextWindow": 204800,
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "pi-hypercharm-provider",
3
- "version": "1.1.1",
3
+ "version": "1.1.3",
4
4
  "description": "HyperCharm provider extension for pi - Access DeepSeek, GLM, Kimi, Qwen, MiniMax, Gemma, and GPT-OSS models through the Charm Hyper API",
5
5
  "type": "module",
6
6
  "main": "index.ts",
package/patch.json CHANGED
@@ -6,7 +6,6 @@
6
6
  "low": null,
7
7
  "medium": null,
8
8
  "high": "high",
9
- "xhigh": "max",
10
9
  "max": "max"
11
10
  },
12
11
  "compat": {
@@ -25,7 +24,6 @@
25
24
  "low": null,
26
25
  "medium": null,
27
26
  "high": "high",
28
- "xhigh": "max",
29
27
  "max": "max"
30
28
  },
31
29
  "compat": {
@@ -174,7 +174,7 @@ function transformModel(apiModel, existingModelsMap) {
174
174
  if (isDeepSeek && isReasoning) {
175
175
  model.maxTokens = 384000;
176
176
  model.thinkingLevelMap = {
177
- minimal: null, low: null, medium: null, high: 'high', xhigh: 'max',
177
+ minimal: null, low: null, medium: null, high: 'high', max: 'max',
178
178
  };
179
179
  model.compat = {
180
180
  thinkingFormat: 'deepseek',