pi-fireworks-provider 1.6.0 → 1.6.2

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.
Files changed (3) hide show
  1. package/README.md +8 -7
  2. package/models.json +17 -0
  3. package/package.json +13 -12
package/README.md CHANGED
@@ -15,7 +15,7 @@ _Kimi, MiniMax, GLM, DeepSeek, GPT-OSS — via Fireworks AI's Anthropic Messages
15
15
 
16
16
  ## Features
17
17
 
18
- - **49+ AI Models** including Kimi K2.5, MiniMax M2.5, GLM 4.5/4.7/5, DeepSeek V3.1/V3.2, DeepSeek V4 Flash, and GPT-OSS
18
+ - **50+ AI Models** including Kimi K2.5, MiniMax M2.5, GLM 4.5/4.7/5, DeepSeek V3.1/V3.2, DeepSeek V4 Flash, and GPT-OSS
19
19
  - **Dual API support** via Fireworks AI's Anthropic Messages and OpenAI-compatible completions endpoints (per-model routing, matching pi core's Fireworks provider)
20
20
  - **Service tiers** — toggle Fireworks `priority` vs `standard` per request on supported models (with priority pricing reflected in cost tracking), via a keybinding, `/fireworks-tier`, and a footer status area
21
21
  - **Preserved thinking** — toggle Fireworks' `reasoning_history: "preserved"` so prior assistant reasoning is retained across turns (better multi-turn recall; uses more tokens), via the `/fireworks-settings` panel, with a model-select notification. Matches neuralwatt/makora's settings-only UX, adapted to Fireworks' single global `reasoning_history` knob
@@ -78,7 +78,8 @@ pi
78
78
  | DeepSeek V4 Flash 0731 | Text | 1.0M | 384K | $0.14 | $0.28 |
79
79
  | DeepSeek V4 Pro | Text | 1.0M | 384K | $1.74 | $3.48 |
80
80
  | DeepSeek V4 Pro (router) | Text | 1.0M | 384K | $1.74 | $3.48 |
81
- | DeepSeek-V4-Pro-0813 | Text | 1.0M | 0 | — | — |
81
+ | DeepSeek V4 Pro 0813 | Text | 1.0M | 384K | — | — |
82
+ | DeepSeek-V4-Flash-Vision-Exp | Text + Image | 1.0M | 0 | — | — |
82
83
  | Gemma 4 26B A4B IT | Text + Image | 262K | 33K | — | — |
83
84
  | Gemma 4 31B IT | Text + Image | 262K | 33K | — | — |
84
85
  | GLM 4.5 | Text | 131K | 131K | $0.55 | $2.19 |
@@ -90,8 +91,8 @@ pi
90
91
  | GLM 5.1 Fast | Text | 203K | 131K | $2.80 | $8.80 |
91
92
  | GLM 5.2 | Text | 1.0M | 131K | $1.40 | $4.40 |
92
93
  | GLM 5.2 Fast | Text | 1.0M | 131K | $2.10 | $6.60 |
93
- | GLM 5.3 Flash | Text + Image | 1.0M | 0 | — | — |
94
- | GLM-5.3 | Text | 1.0M | 0 | — | — |
94
+ | GLM 5.3 | Text | 1.0M | 131K | $1.40 | $4.40 |
95
+ | GLM 5.3 Flash | Text + Image | 1.0M | 131K | $0.15 | $0.50 |
95
96
  | GPT OSS 120B | Text | 131K | 33K | $0.15 | $0.60 |
96
97
  | GPT OSS 20B | Text | 131K | 33K | $0.07 | $0.30 |
97
98
  | Inkling | Text + Image | 1.0M | 262K | $1.00 | $4.05 |
@@ -113,14 +114,14 @@ pi
113
114
  | MiniMax-M2.5 | Text | 197K | 197K | $0.30 | $1.20 |
114
115
  | MiniMax-M2.7 | Text | 197K | 197K | $0.30 | $1.20 |
115
116
  | MiniMax-M3 | Text + Image | 512K | 512K | $0.30 | $1.20 |
116
- | Muse Glimmer 30B | Text + Image | 131K | 0 | — | — |
117
- | Nemotron Lightning 3.5 30B A3B | Text | 262K | 0 | — | — |
117
+ | Muse Glimmer 30B | Text + Image | 131K | 131K | — | — |
118
+ | Nemotron 3.5 Lightning 30B A3B | Text | 262K | 262K | — | — |
118
119
  | NVIDIA Nemotron 3 Ultra NVFP4 | Text | 262K | 66K | $0.60 | $2.40 |
119
120
  | Qwen 3.7 Plus | Text + Image | 262K | 66K | $0.40 | $1.60 |
120
121
  | Qwen3 8B | Text | 41K | 41K | $0.20 | $0.20 |
121
122
  | Qwen3 VL 30B A3B Instruct | Text + Image | 262K | 33K | $0.50 | $0.50 |
122
123
  | Qwen3 VL 30B A3B Thinking | Text + Image | 262K | 33K | $0.50 | $0.50 |
123
- | Qwen3.8-2.4T-A95B | Text | 262K | 0 | — | — |
124
+ | Qwen3.8-2.4T-A95B | Text | 262K | 131K | — | — |
124
125
  *Costs are per million tokens. Prices subject to change - check [fireworks.ai](https://fireworks.ai) for current pricing.*
125
126
 
126
127
  ## Service Tiers
package/models.json CHANGED
@@ -63,6 +63,23 @@
63
63
  "contextWindow": 1048576,
64
64
  "maxTokens": 0
65
65
  },
66
+ {
67
+ "id": "accounts/fireworks/models/deepseek-v4-flash-vision-exp",
68
+ "name": "DeepSeek-V4-Flash-Vision-Exp",
69
+ "reasoning": false,
70
+ "input": [
71
+ "text",
72
+ "image"
73
+ ],
74
+ "cost": {
75
+ "input": 0,
76
+ "output": 0,
77
+ "cacheRead": 0,
78
+ "cacheWrite": 0
79
+ },
80
+ "contextWindow": 1048576,
81
+ "maxTokens": 0
82
+ },
66
83
  {
67
84
  "id": "accounts/fireworks/models/deepseek-v4-pro",
68
85
  "name": "DeepSeek-V4-Pro",
package/package.json CHANGED
@@ -1,9 +1,18 @@
1
1
  {
2
2
  "name": "pi-fireworks-provider",
3
- "version": "1.6.0",
3
+ "version": "1.6.2",
4
4
  "description": "Fireworks AI provider extension for pi - Access Kimi, MiniMax, GLM, DeepSeek, and GPT-OSS models through the Fireworks AI API",
5
5
  "type": "module",
6
6
  "main": "index.ts",
7
+ "scripts": {
8
+ "clean": "echo 'nothing to clean'",
9
+ "build": "echo 'nothing to build'",
10
+ "check": "tsc --noEmit && vitest run && knip --no-gitignore",
11
+ "lint:dead": "knip --no-gitignore",
12
+ "test": "vitest run",
13
+ "test:watch": "vitest",
14
+ "update-models": "node scripts/update-models.js"
15
+ },
7
16
  "keywords": [
8
17
  "pi",
9
18
  "extension",
@@ -32,20 +41,12 @@
32
41
  "@types/node": "^22.20.0",
33
42
  "knip": "6.14.1",
34
43
  "typescript": "6.0.3",
35
- "vitest": "4.1.7"
44
+ "vitest": "4.1.7",
45
+ "@earendil-works/pi-coding-agent": "0.85.1"
36
46
  },
37
47
  "pi": {
38
48
  "extensions": [
39
49
  "./index.ts"
40
50
  ]
41
- },
42
- "scripts": {
43
- "clean": "echo 'nothing to clean'",
44
- "build": "echo 'nothing to build'",
45
- "check": "tsc --noEmit && vitest run && knip --no-gitignore",
46
- "lint:dead": "knip --no-gitignore",
47
- "test": "vitest run",
48
- "test:watch": "vitest",
49
- "update-models": "node scripts/update-models.js"
50
51
  }
51
- }
52
+ }