free-coding-models 0.5.38 → 0.5.40

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.
@@ -0,0 +1,16 @@
1
+ # Changelog v0.5.39 - 2026-06-27
2
+
3
+ ### Added
4
+ - **groq**: +`qwen/qwen3.6-27b` (preview, 131k, A tier)
5
+ - **cloudflare**: +`@cf/zai-org/glm-5.2` (S+, 262k), +`@cf/moonshotai/kimi-k2.5` (S+, 256k)
6
+ - **ovhcloud**: +`Qwen3-Embedding-8B`, +`bge-m3`, +`bge-multilingual-gemma2` (embeddings)
7
+ - **openrouter**: +`cohere/north-mini-code:free` (S tier, 256k), +`nvidia/nemotron-3-ultra-550b-a55b:free` (S+ tier, 1M)
8
+
9
+ ### Changed
10
+ - **openrouter**: full cleanup — removed 11 models no longer on free tier (`qwen3-coder`, `glm-4.5-air`, `kimi-k2.6`, `qwen3-next-80b`, `arcee-trinity`, `hermes-3-405b`, `dolphin-mistral`, `llama-3.3-70b`, `llama-3.2-3b`, `lfm-2.5-1.2b`, `lfm-2.5-thinking`), updated tiers to match live free models page
11
+ - **cloudflare**: fixed `nemotron-3-120b-a12b` context window 32k→128k
12
+ - **openrouter**: corrected ctx for `laguna-xs.2` (131k→262k), `gemma-4-31b` (256k→262k), `gemma-4-26b` (256k→262k), `nemotron-3-super` (262k→1M)
13
+ - **openrouter**: `gemma-4-26b-a4b-it` tier A-→A (corrected)
14
+
15
+ ### Fixed
16
+ - **openrouter**: `nemotron-3-super-120b-a12b:free` ctx 262k→1M (confirmed on openrouter.ai)
@@ -0,0 +1,4 @@
1
+ # Changelog v0.5.40 - 2026-07-07
2
+
3
+ ### Fixed
4
+ - Avoid sending `thinking: { type: 'disabled' }` to Cerebras, Mistral, Groq, and SambaNova during ping probes, preventing immediate 400 errors and redundant retry calls.
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "free-coding-models",
3
- "version": "0.5.38",
3
+ "version": "0.5.40",
4
4
  "description": "Find the fastest coding LLM models in seconds — ping free models from multiple providers, pick the best one for OpenCode, Cursor, or any AI coding assistant.",
5
5
  "keywords": [
6
6
  "nvidia",
package/sources.js CHANGED
@@ -86,6 +86,7 @@ export const groq = [
86
86
  ['openai/gpt-oss-120b', 'GPT OSS 120B', 'S', '60.0%', '131k'],
87
87
  ['openai/gpt-oss-20b', 'GPT OSS 20B', 'A', '42.0%', '131k'],
88
88
  ['qwen/qwen3-32b', 'Qwen3 32B', 'A+', '50.0%', '131k'],
89
+ ['qwen/qwen3.6-27b', 'Qwen3.6 27B', 'A', '-', '131k'],
89
90
  ['groq/compound', 'Groq Compound', 'A', '45.0%', '131k'],
90
91
  ['groq/compound-mini', 'Groq Compound Mini', 'B+', '32.0%', '131k'],
91
92
  ]
@@ -125,38 +126,27 @@ export const sambanova = [
125
126
  // 📖 • Free-tier popular models may be additionally rate-limited by the provider itself during peak hours.
126
127
  // 📖 API keys at https://openrouter.ai/keys
127
128
  export const openrouter = [
128
- // ── S+ tier — live :free chat/coding models ──
129
- ['qwen/qwen3-coder:free', 'Qwen3 Coder 480B', 'S+', '70.6%', '1M'],
130
- ['z-ai/glm-4.5-air:free', 'GLM 4.5 Air', 'S+', '72.0%', '131k'],
129
+ // ── S+ tier — confirmed :free on openrouter.ai/collections/free-models ──
130
+ ['nvidia/nemotron-3-ultra-550b-a55b:free', 'Nemotron 3 Ultra', 'S+', '-', '1M'],
131
131
  ['poolside/laguna-m.1:free', 'Poolside Laguna M.1', 'S+', '-', '262k'],
132
- ['poolside/laguna-xs.2:free', 'Poolside Laguna XS.2','S+', '-', '131k'],
133
- ['moonshotai/kimi-k2.6:free', 'Kimi K2.6', 'S+', '76.8%', '262k'],
134
- // ── S tier — live :free chat/coding models ──
135
- ['qwen/qwen3-next-80b-a3b-instruct:free', 'Qwen3 80B Instruct', 'S', '65.0%', '262k'],
132
+ ['poolside/laguna-xs.2:free', 'Poolside Laguna XS.2','S+', '-', '262k'],
133
+ // ── S tier ──
136
134
  ['openai/gpt-oss-120b:free', 'GPT OSS 120B', 'S', '60.0%', '131k'],
137
- ['arcee-ai/trinity-large-thinking:free', 'Arcee Trinity Large', 'A', '-', '262k'],
138
- ['nvidia/nemotron-3-super-120b-a12b:free', 'Nemotron 3 Super', 'A+', '56.0%', '262k'],
139
- // ── A+ tier — live :free chat/coding models ──
135
+ ['cohere/north-mini-code:free', 'North Mini Code', 'S', '-', '256k'],
136
+ // ── A+ tier ──
137
+ ['nvidia/nemotron-3-super-120b-a12b:free', 'Nemotron 3 Super', 'A+', '56.0%', '1M'],
140
138
  ['nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free', 'Nemotron 3 Omni', 'A+', '52.0%', '256k'],
141
- ['nvidia/nemotron-nano-12b-v2-vl:free', 'Nemotron Nano 12B VL','A', '20.0%', '128k'],
142
139
  ['openrouter/owl-alpha', 'Owl Alpha', 'A+', '-', '1M'],
143
- // ── A tier — live :free chat/coding models ──
144
- ['nousresearch/hermes-3-llama-3.1-405b:free', 'Hermes 3 405B', 'A', '44.0%', '131k'],
140
+ // ── A tier ──
145
141
  ['openai/gpt-oss-20b:free', 'GPT OSS 20B', 'A', '42.0%', '131k'],
146
142
  ['nvidia/nemotron-nano-30b-a3b:free', 'Nemotron Nano 30B', 'A', '43.0%', '256k'],
147
- ['cognitivecomputations/dolphin-mistral-24b-venice-edition:free', 'Dolphin Mistral 24B', 'B+', '30.0%', '33k'],
148
- ['google/gemma-4-31b-it:free', 'Gemma 4 31B', 'A', '45.0%', '256k'],
149
- ['google/gemma-4-26b-a4b-it:free', 'Gemma 4 26B MoE', 'A-', '38.0%', '256k'],
150
- // ── A- tier — confirmed free ──
151
- ['meta-llama/llama-3.3-70b-instruct:free', 'Llama 3.3 70B', 'A-', '39.5%', '131k'],
152
- ['meta-llama/llama-3.2-3b-instruct:free', 'Llama 3.2 3B', 'B', '20.0%', '128k'],
143
+ ['nvidia/nemotron-nano-12b-v2-vl:free', 'Nemotron Nano 12B VL','A', '20.0%', '128k'],
144
+ ['google/gemma-4-31b-it:free', 'Gemma 4 31B', 'A', '45.0%', '262k'],
145
+ ['google/gemma-4-26b-a4b-it:free', 'Gemma 4 26B MoE', 'A', '38.0%', '262k'],
153
146
  // ── B+ tier ──
154
147
  ['nvidia/nemotron-nano-9b-v2:free', 'Nemotron Nano 9B', 'B+', '18.0%', '128k'],
155
148
  // ── B tier ──
156
149
  ['openrouter/free', 'OpenRouter Free', 'B', '-', '200k'],
157
- // ── C tier ──
158
- ['liquid/lfm-2.5-1.2b-instruct:free', 'LFM 2.5 1.2B', 'C', '-', '32k'],
159
- ['liquid/lfm-2.5-1.2b-thinking:free', 'LFM 2.5 Thinking', 'C', '-', '32k'],
160
150
  ]
161
151
 
162
152
  // 📖 GitHub Models source - https://models.github.ai
@@ -270,7 +260,9 @@ export const cloudflare = [
270
260
  ['@cf/zai-org/glm-4.7-flash', 'GLM-4.7-Flash', 'S', '59.2%', '131k'],
271
261
  ['@cf/openai/gpt-oss-120b', 'GPT OSS 120B', 'S', '60.0%', '128k'],
272
262
  // ── A+ tier ──
273
- ['@cf/nvidia/nemotron-3-120b-a12b', 'Nemotron 3 Super', 'A+', '56.0%', '32k'],
263
+ ['@cf/nvidia/nemotron-3-120b-a12b', 'Nemotron 3 Super', 'A+', '56.0%', '128k'],
264
+ ['@cf/zai-org/glm-5.2', 'GLM-5.2', 'S+', '-', '262k'],
265
+ ['@cf/moonshotai/kimi-k2.5', 'Kimi K2.5', 'S+', '-', '256k'],
274
266
  // ── A tier ──
275
267
  ['@cf/meta/llama-4-scout-17b-16e-instruct', 'Llama 4 Scout', 'A', '44.0%', '131k'],
276
268
  ['@cf/mistralai/mistral-7b-instruct-v0.2', 'Mistral 7B v0.2', 'A', '38.0%', '128k'],
@@ -303,6 +295,10 @@ export const ovhcloud = [
303
295
  ['Mistral-7B-Instruct-v0.3', 'Mistral 7B Instruct', 'B', '25.0%', '127k'],
304
296
  ['Mistral-Nemo-Instruct-2407', 'Mistral Nemo', 'B+', '30.0%', '118k'],
305
297
  ['Qwen3.5-9B', 'Qwen3.5 9B', 'B+', '30.0%', '262k'],
298
+ // ── Embeddings ──
299
+ ['Qwen3-Embedding-8B', 'Qwen3 Embedding 8B', 'B', '-', '-'],
300
+ ['bge-m3', 'BGE M3', 'B', '-', '-'],
301
+ ['bge-multilingual-gemma2', 'BGE Multilingual Gemma2','B','-', '-'],
306
302
  // Fix (2026-05-26): Qwen3.5-9B ctx 128k→262k, Mistral-Small ctx 131k→128k, Mistral-Nemo ctx 128k→118k, Mistral-7B ctx 32k→127k
307
303
  ]
308
304
 
package/src/core/ping.js CHANGED
@@ -80,6 +80,7 @@ export function markDisabledThinkingUnsupported(providerKey) {
80
80
  // 📖 shouldUseDisabledThinkingForProvider: central policy for OpenAI-compatible
81
81
  // 📖 probes, shared by regular pings and router health probes.
82
82
  export function shouldUseDisabledThinkingForProvider(providerKey) {
83
+ if (providerKey === 'cerebras' || providerKey === 'mistral' || providerKey === 'groq' || providerKey === 'sambanova') return false
83
84
  return !disabledThinkingUnsupportedProviders.has(providerKey)
84
85
  }
85
86