Models · Conifer

Fig. 01·the frontier

Price against quality

Blended price per million tokens against measured score. Higher and further left is better.

Intelligence and blended price for 223 models
Model	Lab	Intelligence	Price per million tokens	Price frontier	On Conifer	Runs locally
Claude Fable 5	Anthropic	49.6	$20	no	yes	no
Claude Fable 5.1 (max)	Anthropic	53.4	$20	no	yes	no
Claude 4.5 Haiku (Non-reasoning)	Anthropic	15.4	$2	no	yes	no
Claude Opus 4.5 (Non-reasoning)	Anthropic	23.7	$10	no	yes	no
Claude Opus 4.8 (max)	Anthropic	41.8	$10	no	yes	no
Claude Opus 4.6 (Non-reasoning, high)	Anthropic	26.4	$10	no	yes	no
Claude Opus 4.7 (max)	Anthropic	40.7	$10	no	yes	no
Claude Opus 5 (max)	Anthropic	50.8	$10	no	yes	no
Claude Opus 5.5 (max)	Anthropic	57.6	$8	yes	yes	no
Claude 4.5 Sonnet	Anthropic	20.7	$6	no	yes	no
Claude Sonnet 4.6 (Non-reasoning, high)	Anthropic	24.7	$6	no	yes	no
Claude Sonnet 5 (max)	Anthropic	38.2	$4	no	yes	no
Claude Sonnet 5.5 (max)	Anthropic	56	$4	yes	yes	no
Command A	Cohere	7	$4.4	no	yes	no
DeepSeek R1 0528	DeepSeek	13.1	$1.8	no	yes	no
DeepSeek R1 Distill Llama 70B	DeepSeek	7.9	$0.80	no	yes	no
DeepSeek V3.1 (Non-reasoning)	DeepSeek	13.7	$0.84	no	yes	no
DeepSeek V3.1 Terminus	DeepSeek	14.8	$1.9	no	yes	no
DeepSeek V3.2	DeepSeek	21.5	$0.32	no	yes	no
DeepSeek V4 Flash 0731 (max)	DeepSeek	34.3	$0.66	no	yes	no
DeepSeek V4 Flash Vision (max)	DeepSeek	34.8	$0.66	no	yes	no
DeepSeek V4 Pro 0813 (max)	DeepSeek	36	$2.0	no	yes	no
DeepSeek V4.1 Flash (max)	DeepSeek	39.5	$0.53	no	yes	no
Gemini 2.5 Flash	Google	13.1	$0.85	no	yes	no
Gemini 2.5 Flash-Lite (Non-reasoning)	Google	6.7	$0.17	no	yes	no
Gemini 2.5 Pro	Google	16.1	$3.4	no	yes	no
Gemini 3 Flash (Non-reasoning)	Google	17.9	$1.1	no	yes	no
Gemini 3.1 Flash-Lite	Google	15.6	$0.56	no	yes	no
Gemini 3.1 Pro Preview	Google	29.7	$4.5	no	yes	no
Gemini 3.5 Flash	Google	32.6	$3.4	no	yes	no
Gemini 3.5 Flash-Lite	Google	22.2	$0.85	no	yes	no
Gemini 3.6 Flash	Google	34	$1.5	no	yes	no
Gemini 3.7 Flash (high)	Google	39.1	$1.5	no	yes	no
Gemini 3.8 Flash (high)	Google	40.9	$1.5	no	yes	no
Gemma 4 26B A4B	Google	16.7	$0.17	no	yes	no
GLM-4.5-Air	Z AI	11.1	$0.37	no	yes	no
GLM-4.5V	Z AI	7.6	$0.90	no	yes	no
GLM-4.6	Z AI	18.5	$0.96	no	yes	no
GLM-4.6V	Z AI	11.2	$0.45	no	yes	no
GLM-4.7	Z AI	22.2	$1	no	yes	no
GLM-4.7-Flash	Z AI	14.9	$0.15	no	yes	no
GLM-5	Z AI	27.9	$1.6	no	yes	no
GLM-5.1	Z AI	26.1	$2.0	no	yes	no
GLM-5.2 (max)	Z AI	33.7	$2.1	no	yes	no
GLM-5.3 (max)	Z AI	44.8	$2.1	no	yes	no
GLM-5.3-Flash	Z AI	41.8	$0.24	yes	yes	no
GPT-3.5 Turbo	OpenAI	5.5	$0.75	no	yes	no
GPT-4.1	OpenAI	12.7	$3.5	no	yes	no
GPT-4.1 mini	OpenAI	10.2	$0.70	no	yes	no
GPT-4.1 nano	OpenAI	7.8	$0.17	no	yes	no
GPT-4o (May)	OpenAI	7.3	$7.5	no	yes	no
GPT-4o (Aug)	OpenAI	7.7	$4.4	no	yes	no
GPT-4o (Nov)	OpenAI	8.4	$4.4	no	yes	no
GPT-4o mini	OpenAI	6.7	$0.26	no	yes	no
GPT-5 (high)	OpenAI	23	$3.4	no	yes	no
GPT-5 mini (high)	OpenAI	16.8	$0.69	no	yes	no
GPT-5 nano (high)	OpenAI	13	$0.14	no	yes	no
GPT-5.1 (high)	OpenAI	24.7	$3.4	no	yes	no
GPT-5.2 (xhigh)	OpenAI	30.4	$4.8	no	yes	no
GPT-5.3 Codex (xhigh)	OpenAI	32.5	$4.8	no	yes	no
GPT-5.4 (xhigh)	OpenAI	39	$5.6	no	yes	no
GPT-5.4 mini (xhigh)	OpenAI	24.1	$1.7	no	yes	no
GPT-5.4 nano (xhigh)	OpenAI	20.7	$0.46	no	yes	no
GPT-5.5 (xhigh)	OpenAI	38.4	$11	no	yes	no
GPT-5.6 Luna (max)	OpenAI	37.3	$0.45	no	yes	no
GPT-5.6 Sol (max)	OpenAI	47	$8	no	yes	no
GPT-5.6 Terra (max)	OpenAI	42.1	$4.5	no	yes	no
GPT-6 Astra (max)	OpenAI	52.7	$20	no	yes	no
GPT-6 Luna (max)	OpenAI	37.3	$0.20	no	yes	no
GPT-6 Sol (max)	OpenAI	47.5	$4	no	yes	no
GPT-6.1 Sol (max)	OpenAI	51.8	$4	no	yes	no
gpt-oss-120b (high)	OpenAI	11.6	$0.26	no	yes	no
gpt-oss-20b (high)	OpenAI	9	$0.10	no	yes	no
Granite 4.2 8B	IBM	11.1	$0.11	no	yes	no
Grok 4.20 0309 (Non-reasoning)	SpaceXAI	14.6	$3	no	yes	no
Grok 4.20 0309	SpaceXAI	25.2	$3	no	yes	no
Grok 4.3 (high)	SpaceXAI	24.9	$1.6	no	yes	no
Grok 4.5 (high)	SpaceXAI	38.8	$3	no	yes	no
Grok 4.6 (high)	SpaceXAI	44.3	$3	no	yes	no
Grok 4.7 (xhigh)	SpaceXAI	46.4	$3	no	yes	no
Grok Build 0.1 0616	SpaceXAI	27.2	$1.3	no	yes	no
Hermes 3 - Llama-3.1 70B	Nous Research	6	$0.70	no	yes	no
Hermes 4 405B	Nous Research	7.5	$1.5	no	yes	no
Hy3	Tencent	25.3	$0.24	no	yes	no
Inkling	Thinking Machines	25	$1.8	no	yes	no
Inkling Small	Thinking Machines	27.8	$0.53	no	yes	no
Kimi K2 0905	Kimi	15.3	$1.1	no	yes	no
Kimi K2 Thinking	Kimi	22	$1.1	no	yes	no
Kimi K2.5	Kimi	23.5	$1.1	no	yes	no
Kimi K2.6	Kimi	27	$1.7	no	yes	no
Kimi K2.7 Code	Kimi	25.8	$1.7	no	yes	no
Kimi K3 (max)	Kimi	43.6	$6	no	yes	no
Ling 3.0 Flash	InclusionAI	20.1	$0.11	no	yes	no
Ling-3.0-flash-VL	InclusionAI	24.6	$0.11	yes	yes	no
Llama 3.1 70B	Meta	6.6	$0.56	no	yes	no
Llama 3.1 8B	Meta	6.9	$0.03	yes	yes	no
Llama 3.3 70B	Meta	7.7	$0.71	no	yes	no
Llama 4 Maverick	Meta	10	$0.42	no	yes	no
Llama 4 Scout	Meta	8.1	$0.31	no	yes	no
Mercury 2	Inception	13.8	$0.38	no	yes	no
MiMo-V2.5	Xiaomi	25.2	$0.17	no	yes	no
MiMo-V2.5-Pro	Xiaomi	26	$0.54	no	yes	no
MiniMax-M2	MiniMax	18.6	$0.53	no	yes	no
MiniMax-M2.1	MiniMax	20.9	$0.53	no	yes	no
MiniMax-M2.5	MiniMax	22.8	$0.53	no	yes	no
MiniMax-M2.7	MiniMax	22.8	$0.53	no	yes	no
MiniMax-M3	MiniMax	29.2	$0.53	no	yes	no
Ministral 3 14B	Mistral	6	$0.20	no	yes	no
Ministral 3 3B	Mistral	4.8	$0.10	no	yes	yes
Ministral 3 8B	Mistral	5.5	$0.15	no	yes	no
Mistral Large 3	Mistral	9.3	$0.75	no	yes	no
Mistral Medium 3	Mistral	9	$0.80	no	yes	no
Mistral Medium 3.5	Mistral	14.2	$3	no	yes	no
Mistral Small 3	Mistral	6.7	$0.06	no	yes	no
Mistral Small 4	Mistral	11.3	$0.26	no	yes	no
Muse Glimmer (high)	Meta	17.5	$0.58	no	yes	no
Muse Spark 1.3 (max)	Meta	48.1	$2	yes	yes	no
Nemotron 3 Nano	NVIDIA	8.9	$0.09	no	yes	no
Nemotron 3 Super	NVIDIA	12.8	$0.45	no	yes	no
Nemotron 3 Ultra	NVIDIA	22.9	$1.1	no	yes	no
Nemotron 3.5 Lightning	NVIDIA	12.9	$0.11	no	yes	no
o1	OpenAI	15.2	$26	no	yes	no
o1-pro	OpenAI	12.4	$263	no	yes	no
o3	OpenAI	20.2	$3.5	no	yes	no
o3-mini	OpenAI	12.5	$1.9	no	yes	no
o4-mini (high)	OpenAI	16.7	$1.9	no	yes	no
Phi-4	Microsoft	5.9	$0.22	no	yes	no
Qwen2.5 72B	Alibaba	7.7	$0.48	no	yes	no
Qwen3 14B	Alibaba	8.2	$1.3	no	yes	no
Qwen3 235B A22B 2507	Alibaba	12.7	$0.75	no	yes	no
Qwen3 30B A3B 2507	Alibaba	9.8	$0.75	no	yes	no
Qwen3 30B A3B 2507 (Non-reasoning)	Alibaba	7.5	$0.35	no	yes	no
Qwen3 32B	Alibaba	8.6	$0.28	no	yes	no
Qwen3 Coder 30B A3B	Alibaba	9.6	$0.90	no	yes	no
Qwen3 Coder 480B	Alibaba	11.9	$3	no	yes	no
Qwen3 Coder Next	Alibaba	9.2	$0.56	no	yes	no
Qwen3 Max	Alibaba	15.6	$2.4	no	yes	no
Qwen3 Next 80B A3B (Reasoning)	Alibaba	11.2	$0.41	no	yes	no
Qwen3 VL 235B A22B	Alibaba	9.9	$0.70	no	yes	no
Qwen3 VL 235B A22B (Reasoning)	Alibaba	13.4	$1.3	no	yes	no
Qwen3 VL 30B A3B	Alibaba	7.9	$0.35	no	yes	no
Qwen3.5 122B A10B	Alibaba	15.6	$1.1	no	yes	no
Qwen3.5 27B	Alibaba	22.9	$0.82	no	yes	yes
Qwen3.5 35B A3B	Alibaba	19.3	$0.69	no	yes	no
Qwen3.5 397B A17B	Alibaba	18.4	$1.4	no	yes	no
Qwen3.5 9B	Alibaba	11.2	$0.15	no	yes	yes
Qwen3.6 27B	Alibaba	21.4	$1.4	no	yes	yes
Qwen3.6 35B A3B	Alibaba	18.2	$0.84	no	yes	no
Qwen3.6 Plus	Alibaba	27	$1.1	no	yes	no
Qwen3.7 Max	Alibaba	29.5	$3.8	no	yes	no
Qwen3.7 Plus	Alibaba	25.2	$0.70	no	yes	no
Qwen3.8 2.4T A95B	Alibaba	39.9	$3	no	yes	no
Qwen3.8 27B (xhigh)	Alibaba	33.7	$1.1	no	yes	yes
Qwen3.8-Flash-Next	Alibaba	39.8	$0.23	yes	yes	no
Qwen3.8 Max (0902)	Alibaba	45.4	$3	no	yes	no
Reka Flash 3	Reka AI	5.6	$0.35	no	yes	no
Solar Pro 3	Upstage	7.8	$0.26	no	yes	no
Solar Pro 4	Upstage	28.2	$0.53	no	yes	no
Step 3.7 Flash	StepFun	19.5	$0.44	no	yes	no
Qwen3.5 Omni Plus	Alibaba	20.4	$1.5	no	no	no
Nova 2.0 Omni (medium)	Amazon	13.6	$0.85	no	no	no
Qwen3.5 122B A10B (Non-reasoning)	Alibaba	17.7	$1.1	no	no	no
Gemma 4 12B (Non-reasoning)	Google	9.4	$0.15	no	no	no
Apodex 1.1	Apodex	26.4	$0.97	no	no	no
Step 5 Preview	StepFun	43.7	$1.4	no	no	no
Qwen3 Omni 30B A3B	Alibaba	6	$0.43	no	no	no
Granite 4.2 3B	IBM	9.1	$0.05	yes	no	no
Nova 2.0 Pro Preview (medium)	Amazon	14.2	$3.4	no	no	no
Granite 4.2 30B	IBM	14.8	$0.28	no	no	no
MiMo-V2.6-Flash	Xiaomi	37.9	$0.17	yes	no	no
Mistral Small 4 (Non-reasoning)	Mistral	9	$0.26	no	no	no
Nova 2.0 Omni (Non-reasoning)	Amazon	8.2	$0.85	no	no	no
Qwen3.5 9B (Non-reasoning)	Alibaba	13.3	$0.19	no	no	no
Nemotron 3 Nano (Non-reasoning)	NVIDIA	6.8	$0.09	no	no	no
MiMo-V2.6-Pro	Xiaomi	46.3	$0.54	yes	no	no
Qwen3 Next 80B A3B	Alibaba	9.6	$0.41	no	no	no
NVIDIA Nemotron Nano 9B V2 (Non-reasoning)	NVIDIA	6.8	$0.09	no	no	no
Qwen3.5 4B (Non-reasoning)	Alibaba	10.8	$0.06	no	no	no
Qwen3.5 397B A17B (Non-reasoning)	Alibaba	21.4	$1.4	no	no	no
Nova Micro	Amazon	5.9	$0.06	no	no	no
Qwen3.8 27B	Alibaba	20.2	$1.1	no	no	yes
Sarvam 105B (high)	Sarvam	8.8	$0.07	no	no	no
Ring-flash-2.0	InclusionAI	7.2	$0.25	no	no	no
Apertus 70B Instruct	Swiss AI Initiative	5.1	$1.3	no	no	no
Gemma 4 E4B (Non-reasoning)	Google	7.5	$0.04	no	no	no
LongCat 2.0	LongCat	19.1	$0.53	no	no	no
Granite 4.0 H Small	IBM	6	$0.11	no	no	no
Llama 3.2 11B (Vision)	Meta	5.4	$0.34	no	no	no
Nova 2.0 Lite (high)	Amazon	13.4	$0.85	no	no	no
GPT-6 Luna (Non-reasoning)	OpenAI	18.3	$0.20	no	no	no
Qwen3.5 4B	Alibaba	13.1	$0.06	yes	no	yes
Qwen3.6 35B A3B (Non-reasoning)	Alibaba	15.2	$0.84	no	no	no
Ling-3.0-flash-Fin	InclusionAI	22.6	$0.11	no	no	no
Qwen3.5 Omni Flash	Alibaba	12.5	$0.28	no	no	no
Magistral Small 1.2	Mistral	8.6	$0.75	no	no	no
Gemma 4 26B A4B (Non-reasoning)	Google	13.1	$0.20	no	no	no
DeepSeek V4.1 Flash (Non-reasoning)	DeepSeek	24.7	$0.53	no	no	no
Sarvam 30B (high)	Sarvam	6.6	$0.05	no	no	no
Mercury 2.5	Inception	12.3	$0.38	no	no	no
Nova 2.0 Lite (Non-reasoning)	Amazon	8.7	$0.85	no	no	no
Celeris-1	Celeris	6.3	$0.33	no	no	no
Apertus 8B Instruct	Swiss AI Initiative	4.8	$0.13	no	no	no
Nemotron 3 Nano Omni 30B A3B	NVIDIA	10.3	$0.42	no	no	no
Olmo 3 7B	Allen Institute for AI	5.2	$0.13	no	no	no
GPT-5.6 Terra (Non-reasoning)	OpenAI	20.8	$4.5	no	no	no
Nova 2.0 Pro Preview (Non-reasoning)	Amazon	10	$3.4	no	no	no
Claude 4.5 Haiku	Anthropic	16.9	$2	no	no	no
NVIDIA Nemotron Nano 12B v2 VL (Non-reasoning)	NVIDIA	5.8	$0.30	no	no	no
ERNIE 4.5 300B A47B	Baidu	7.5	$0.48	no	no	no
Nova 2.0 Omni (low)	Amazon	11.1	$0.85	no	no	no
Trinity Large Thinking	Arcee AI	10.8	$0.39	no	no	no
Qwen3 Omni 30B A3B (Reasoning)	Alibaba	7.8	$0.43	no	no	no
GPT-5.5 Instant (June 2026)	OpenAI	26	$11	no	no	no
Gemma 4 31B (Non-reasoning)	Google	13.9	$0.20	no	no	no
Hermes 4 405B (Non-reasoning)	Nous Research	7.4	$1.5	no	no	no
GPT-6 Sol (Non-reasoning)	OpenAI	28.1	$4	no	no	no
Gemma 4 E4B	Google	8.9	$0.04	yes	no	yes
Ring-2.6-1T	InclusionAI	16.6	$0.85	no	no	no
NVIDIA Nemotron Nano 9B V2	NVIDIA	7.4	$0.07	no	no	no
Nova 2.0 Pro Preview (low)	Amazon	12.8	$3.4	no	no	no
Quasar 438B (max)	Multiverse Computing	26.7	$0.90	no	no	no
KAT-Coder-Pro V2	KwaiKAT	21.7	$0.53	no	no	no
Gemma 4 12B	Google	14.2	$0.15	no	no	yes

02·release dates

Intelligence by release date

Claude Opus 5.5maxAnthropicreleased 2026-09-22intelligence 57.6on Conifer $4 in · $20 out

03·running cost

Cost per hour of a steady coding workload

the meters start when this panel is on screen

DeepSeek V4.1 Flash (max)DeepSeek$1.4/hour$1.4/hour · $0.30 in / $1.2 out per Mtok
Gemini 3.8 Flash (high)Google$3.6/hour$3.6/hour · $0.75 in / $3.8 out per Mtok
GLM-5.3 (max)Z AI$6.1/hour$6.1/hour · $1.4 in / $4.4 out per Mtok
Grok 4.7 (xhigh)SpaceXAI$8.6/hour$8.6/hour · $2 in / $6 out per Mtok
Qwen3.8 Max (0902)Alibaba$8.6/hour$8.6/hour · $2 in / $6 out per Mtok
Kimi K3 (max)Kimi$14/hour$14/hour · $3 in / $15 out per Mtok
Claude Opus 5.5 (max)Anthropic$19/hour$19/hour · $4 in / $20 out per Mtok
GPT-6 Astra (max)OpenAI$43/hour$43/hour · $9 in / $45 out per Mtok

04·time to first token

Time to first token, best model per lab

Median time to first token; the reply then streams at each model's output speed.

8.84s

In one sentence, what is a Pareto frontier?

DeepSeek V4.1 Flash (max)
A Pareto frontier is the set of choices where nothing can be improved without making something else worse.
first token at 0.94s
Claude Opus 4.6 (Non-reasoning, high)
A Pareto frontier is the set of choices where nothing can be improved without making something else worse.
first token at 2.0s
GLM-5.3 (max)
A Pareto frontier is the set of choices where nothing can be improved without making something else worse.
first token at 2.8s
Qwen3.8 Max (0902)
A Pareto frontier is the set of choices where nothing can be improved without making something else worse.
first token at 3.0s
Kimi K3 (max)
A Pareto frontier is the set of choices where nothing can be improved without making something else worse.
first token at 3.0s
o3
A Pareto frontier is the set of choices where nothing can be improved without making something else worse.
first token at 5.8s
Grok 4.5 (high)
A Pareto frontier is the set of choices where nothing can be improved without making something else worse.
first token at 7.1s
Gemini 3.5 Flash-Lite
A Pareto frontier is the set of choices where nothing can be improved without making something else worse.
first token at 8.8s

05·hallucination rate

Wrong answers instead of declining

Of 100 questions a model could not answer correctly, how many it answered anyway instead of declining.

29 of 100Qwen3.8 Max (0902)guessed · 71 declined
29 of 100Grok 4.7 (xhigh)guessed · 71 declined
30 of 100GLM-5.3 (max)guessed · 70 declined
51 of 100GPT-6 Astra (max)guessed · 49 declined
53 of 100Kimi K3 (max)guessed · 47 declined
55 of 100Gemini 3.8 Flash (high)guessed · 45 declined
59 of 100Claude Opus 5.5 (max)guessed · 41 declined
97 of 100DeepSeek V4.1 Flash (max)guessed · 3 declined

Qwen3.8 Max (0902)Alibabaintelligence 45.4hallucinated 28.8%on Conifer $2 in · $6 out

06·labs

Price against intelligence, by lab

Intelligence Index against blended list price per million tokens, log scale, the same axes in every panel.

Claude Opus 5.5maxAnthropicintelligence 57.6blended $8/Mtokon Conifer $4 in · $20 outserved by Conifer

07·model category

early access · 15 Sep 2026

Structured decisions

Jev returns typed decisions with probabilities, rather than generated text. TypeSafe’s own workflow evaluation, checked 20 Sep 2026.

Agreement with reference labels
% · higher is better

Cost per workflow
USD · lower is better · log scale

Time per workflow
seconds · lower is better · log scale

Vendor-reported workflow runs, default reasoning; the mean weights four workflows equally. Reference labels come from GPT-6 Astra and Claude Fable 5.1 at high thinking. Agreement is not the AA Intelligence Index; elapsed workflow time is not token speed or time to first token.

Conifer price: $0.042 input / $0.00 output per million tokens. Listed in Conifer’s catalog for structured decisions. Type safety constrains output shape, not factual correctness. Release and pricing · Evaluation and methodology. Figures retain the source’s rounding; model names retain its labels.

08·output speed

The same paragraph at six output speeds

Each model types at its median measured output speed.

Mercury 2777 tok/s

Conifer routes each request to a model that fits it. A small local model takes a quick question and a larger hosted model takes a hard one, and nothing leaves the machine unless the route says it should. The runtime installs with one command, the router reads the request before it picks a model, and every answer carries a receipt that says what it cost. Your own hardware and your own keys cost nothing extra, and cloud tokens are billed at the provider's own rate. One key and one endpoint reach every model, and the response looks the same whether it came from the laptop or from a hosted model.

0.2s

Muse Spark 1.3168 tok/s

Conifer routes each request to a model that fits it. A small local model takes a quick question and a larger hosted model takes a hard one, and nothing leaves the machine unless the route says it should. The runtime installs with one command, the router reads the request before it picks a model, and every answer carries a receipt that says what it cost. Your own hardware and your own keys cost nothing extra, and cloud tokens are billed at the provider's own rate. One key and one endpoint reach every model, and the response looks the same whether it came from the laptop or from a hosted model.

0.9s

Claude Opus 5.592 tok/s

Conifer routes each request to a model that fits it. A small local model takes a quick question and a larger hosted model takes a hard one, and nothing leaves the machine unless the route says it should. The runtime installs with one command, the router reads the request before it picks a model, and every answer carries a receipt that says what it cost. Your own hardware and your own keys cost nothing extra, and cloud tokens are billed at the provider's own rate. One key and one endpoint reach every model, and the response looks the same whether it came from the laptop or from a hosted model.

1.6s

GPT-6 Astra57 tok/s

Conifer routes each request to a model that fits it. A small local model takes a quick question and a larger hosted model takes a hard one, and nothing leaves the machine unless the route says it should. The runtime installs with one command, the router reads the request before it picks a model, and every answer carries a receipt that says what it cost. Your own hardware and your own keys cost nothing extra, and cloud tokens are billed at the provider's own rate. One key and one endpoint reach every model, and the response looks the same whether it came from the laptop or from a hosted model.

2.6s

Hermes 3 - Llama-3.1 70B29 tok/s

Conifer routes each request to a model that fits it. A small local model takes a quick question and a larger hosted model takes a hard one, and nothing leaves the machine unless the route says it should. The runtime installs with one command, the router reads the request before it picks a model, and every answer carries a receipt that says what it cost. Your own hardware and your own keys cost nothing extra, and cloud tokens are billed at the provider's own rate. One key and one endpoint reach every model, and the response looks the same whether it came from the laptop or from a hosted model.

5.2s

09·memory footprint

What fits on your machine

The next size up, Mistral Small 3.2 24B Instruct 2506, needs about 15 GB.

10·served models

197 served models

Every served model

Input and output prices are USD per million tokens, what Conifer charges; a struck figure is the lab's list price. Decision models are labeled; dots mark unpublished or inapplicable measurements.


Claude Opus 5.5max	Anthropic	closed	57.6	92	685s	1M	$4	$20	2026-09-22
Claude Sonnet 5.5max	Anthropic	closed	56	137	379s	1M	$2	$10	2026-09-28
Claude Fable 5.1max10% off	Anthropic	closed	53.4	68	276s	1M	~~$10~~ $9	~~$50~~ $45	2026-09-01
GPT-6 Astramax10% off	OpenAI	closed	52.7	57	306s	1.1M	~~$10~~ $9	~~$50~~ $45	2026-09-03
GPT-6.1 Solmax	OpenAI	closed	51.8	67	268s	1.1M	$2	$10	2026-09-29
Claude Opus 5max	Anthropic	closed	50.8	57	59s	1M	$5	$25	2026-07-24
Claude Fable 5	Anthropic	closed	49.6	63	91s	1M	$10	$50	2026-06-09
Muse Spark 1.3max	Meta	closed	48.1	168	29s	1M	$1.3	$4.3	2026-09-02
GPT-6 Solmax	OpenAI	closed	47.5	76	197s	1.1M	$2	$10	2026-09-22
GPT-5.6 Solmax	OpenAI	closed	47	82	99s	1.1M	$4	$20	2026-07-09
Grok 4.7xhigh	SpaceXAI	closed	46.4	74	55s	500k	$2	$6	2026-09-21
Qwen3.8 Max (0902)	Alibaba	closed	45.4	39	3.0s	256k	$2	$6	2026-09-02
GLM-5.3max	Z AI	open	44.8	83	2.8s	1M	$1.4	$4.4	2026-08-18
Grok 4.6high	SpaceXAI	closed	44.3	67	35s	500k	$2	$6	2026-08-12
Kimi K3max	Kimi	open	43.6	39	3.0s	1M	$3	$15	2026-07-16
GPT-5.6 Terramax	OpenAI	closed	42.1	95	164s	1.1M	$2	$12	2026-07-09
Claude Opus 4.8max	Anthropic	closed	41.8	53	50s	1M	$5	$25	2026-05-28
GLM-5.3-Flash	Z AI	open	41.8	45	3.4s	1M	$0.15	$0.50	2026-08-26
Gemini 3.8 Flashhigh	Google	closed	40.9	242	19s	1M	$0.75	$3.8	2026-09-02
Claude Opus 4.7max	Anthropic	closed	40.7	46	24s	1M	$5	$25	2026-04-16
Qwen3.8 2.4T A95B	Alibaba	open	39.9	39	2.9s	262k	$2	$6	2026-08-12
Qwen3.8-Flash-Next	Alibaba	open	39.8	58	2.5s	1M	$0.15	$0.47	2026-08-26
DeepSeek V4.1 Flashmax	DeepSeek	open	39.5	212	0.94s	1M	$0.30	$1.2	2026-09-10
Gemini 3.7 Flashhigh	Google	closed	39.1	290	15s	1M	$0.75	$3.8	2026-08-13
GPT-5.4xhigh	OpenAI	closed	39	103	158s	1.1M	$2.5	$15	2026-03-05
Grok 4.5high	SpaceXAI	closed	38.8	58	7.1s	500k	$2	$6	2026-07-08
GPT-5.5xhigh	OpenAI	closed	38.4	94	44s	1.1M	$5	$30	2026-04-23
Claude Sonnet 5max	Anthropic	closed	38.2	75	197s	1M	$2	$10	2026-06-30
GPT-5.6 Lunamax	OpenAI	closed	37.3	118	104s	1.1M	$0.20	$1.2	2026-07-09
GPT-6 Lunamax	OpenAI	closed	37.3	149	97s	1.1M	$0.10	$0.50	2026-09-22
DeepSeek V4 Pro 0813max	DeepSeek	open	36	81	1.7s	1M	$1.3	$2.6	2026-08-13
DeepSeek V4 Flash Visionmax	DeepSeek	closed	34.8	227	1.1s	1M	$0.44	$1.3	2026-08-21
DeepSeek V4 Flash 0731max	DeepSeek	open	34.3	218	1.0s	1M	$0.30	$1.2	2026-07-31
Gemini 3.6 Flash	Google	closed	34	184	18s	1M	$0.75	$3.8	2026-07-21
GLM-5.2max	Z AI	open	33.7	79	3.1s	1M	$1.4	$4.4	2026-06-16
Qwen3.8 27Bxhigh	Alibaba	open	33.7	44	3.9s	262k	$0.50	$3	2026-08-14
Gemini 3.5 Flash	Google	closed	32.6	195	17s	1M	$1.5	$9	2026-05-19
GPT-5.3 Codexxhigh	OpenAI	closed	32.5	97	69s	400k	$1.8	$14	2026-02-05
GPT-5.2xhigh	OpenAI	closed	30.4	68	120s	400k	$1.8	$14	2025-12-11
Gemini 3.1 Pro Preview	Google	closed	29.7	113	26s	1M	$2	$12	2026-02-19
Qwen3.7 Max	Alibaba	closed	29.5	202	2.3s	1M	$2.5	$7.5	2026-05-19
MiniMax-M3	MiniMax	open	29.2	108	1.5s	1M	$0.30	$1.2	2026-06-01
Solar Pro 4	Upstage	closed	28.2	83	2.0s	524k	$0.03	$0.12	2026-08-06
GLM-5	Z AI	open	27.9	78	1.5s	203k	$1	$3.2	2026-02-11
Inkling Small	Thinking Machines	open	27.8	208	2.0s	524k	$0.50	$1.2	2026-07-30
Grok Build 0.1 0616	SpaceXAI	closed	27.2	41	0.62s	256k	$1	$2	·
Kimi K2.6	Kimi	open	27	72	3.0s	262k	$0.95	$4	2026-04-20
Qwen3.6 Plus	Alibaba	closed	27	56	2.1s	1M	$0.50	$3	·
GLM-5-Turbo	Z AI	closed	26.6	·	·	203k	$1.2	$4	2026-03-15
Claude Opus 4.6 (Non-reasoning, high)	Anthropic	closed	26.4	37	2.0s	1M	$5	$25	2026-02-05
GLM-5.1	Z AI	open	26.1	65	1.7s	203k	$1.1	$3.5	2026-04-07
MiMo-V2.5-Pro	Xiaomi	open	26	32	2.4s	1M	$1	$3	2026-04-22
Kimi K2.7 Code	Kimi	open	25.8	81	2.7s	262k	$0.95	$4	2026-06-12
Hy3	Tencent	open	25.3	90	3.0s	262k	$0.14	$0.58	2026-07-06
Grok 4.20 0309	SpaceXAI	closed	25.2	·	·	1M	$1.3	$2.5	·
MiMo-V2.5	Xiaomi	open	25.2	45	25s	262k	$0.40	$2	2026-04-22
Qwen3.7 Plus	Alibaba	closed	25.2	57	2.1s	1M	$0.32	$1.3	2026-06-01
Inkling	Thinking Machines	open	25	177	1.8s	524k	$1	$4.0	2026-07-15
Grok 4.3high	SpaceXAI	closed	24.9	111	30s	1M	$1.3	$2.5	2026-04-30
Claude Sonnet 4.6 (Non-reasoning, high)	Anthropic	closed	24.7	43	1.4s	1M	$3	$15	2026-02-17
GPT-5.1high	OpenAI	closed	24.7	112	38s	400k	$1.3	$10	2025-11-13
Ling-3.0-flash-VL	InclusionAI	open	24.6	142	1.9s	131k	$0.06	$0.18	·
GPT-5.4 minixhigh	OpenAI	closed	24.1	224	128s	400k	$0.75	$4.5	2026-03-17
Claude Opus 4.5 (Non-reasoning)	Anthropic	closed	23.7	45	1.2s	200k	$5	$25	·
GLM 5V Turbo	Z AI	closed	23.5	·	·	205k	$1.2	$4	2026-04-01
Kimi K2.5	Kimi	open	23.5	·	·	262k	$0.60	$3	2026-01-27
GPT-5high	OpenAI	closed	23	95	66s	400k	$1.3	$10	2025-08-07
Nemotron 3 Ultra	NVIDIA	open	22.9	214	1.0s	262k	$0.50	$2.2	2026-06-04
Qwen3.5 27B	Alibaba	open	22.9	76	5.7s	262k	$0.26	$2.6	2026-02-24
MiniMax-M2.5	MiniMax	open	22.8	84	1.5s	205k	$0.30	$1.2	2026-02-12
MiniMax-M2.7	MiniMax	open	22.8	59	1.6s	197k	$0.30	$1.2	2026-03-18
Gemini 3.5 Flash-Lite	Google	closed	22.2	320	8.8s	1M	$0.30	$2.5	2026-07-21
GLM-4.7	Z AI	open	22.2	87	1.2s	200k	$0.60	$2.2	2025-12-22
Kimi K2 Thinking	Kimi	open	22	124	1.4s	262k	$0.60	$2.5	·
DeepSeek V3.2	DeepSeek	open	21.5	·	·	164k	$0.26	$0.38	2025-12-01
Qwen3.6 27B	Alibaba	open	21.4	56	3.7s	262k	$0.32	$3.2	2026-04-22
Qwen3 Max Thinking	Alibaba	closed	21.3	·	·	256k	$1.2	$6	2026-01-26
MiniMax-M2.1	MiniMax	open	20.9	91	1.5s	205k	$0.30	$1.2	2025-12-23
Claude 4.5 Sonnet	Anthropic	closed	20.7	40	13s	200k	$3	$15	·
GPT-5.4 nanoxhigh	OpenAI	closed	20.7	164	83s	400k	$0.20	$1.3	2026-03-17
o3	OpenAI	closed	20.2	136	5.8s	200k	$2	$8	2025-04-16
Ling 3.0 Flash	InclusionAI	open	20.1	331	3.1s	131k	$0.06	$0.18	2026-08-04
Step 3.7 Flash	StepFun	open	19.5	183	2.6s	262k	$0.20	$1.1	2026-05-29
Qwen3.5 35B A3B	Alibaba	open	19.3	147	2.1s	262k	$0.14	$1	2026-02-24
Gemma 4 31B	Google	open	19	36	1.0s	128k	$0.40	$0.60	2026-04-02
MiniMax-M2	MiniMax	open	18.6	86	1.4s	205k	$0.30	$1.2	2025-10-26
GLM-4.6	Z AI	open	18.5	31	4.5s	203k	$0.50	$2	2025-09-30
Qwen3.5 397B A17B	Alibaba	open	18.4	89	2.3s	262k	$0.45	$3	2026-02-16
Qwen3.6 35B A3B	Alibaba	open	18.2	124	2.1s	262k	$0.10	$0.95	2026-04-16
Gemini 3 Flash (Non-reasoning)	Google	closed	17.9	170	1.1s	1M	$0.50	$3	2025-12-17
Muse Glimmerhigh	Meta	open	17.5	144	1.4s	131k	$0.30	$1.2	2026-08-10
GPT-5 minihigh	OpenAI	closed	16.8	113	54s	400k	$0.25	$2	2025-08-07
Gemma 4 26B A4B	Google	open	16.7	·	·	262k	$0.07	$0.34	2026-04-02
o4-minihigh	OpenAI	closed	16.7	141	22s	200k	$1.1	$4.4	2025-04-16
Gemini 2.5 Pro	Google	closed	16.1	124	22s	1M	$1.3	$10	2025-06-05
Gemini 3.1 Flash-Lite	Google	closed	15.6	266	5.8s	1M	$0.25	$1.5	2026-03-03
Qwen3 Max	Alibaba	closed	15.6	55	2.2s	256k	$1.2	$6	2025-09-23
Qwen3.5 122B A10B	Alibaba	open	15.6	126	2.4s	262k	$0.29	$2.4	2026-02-24
Claude 4.5 Haiku (Non-reasoning)	Anthropic	closed	15.4	83	0.63s	200k	$1	$5	2025-10-15
Kimi K2 0905	Kimi	open	15.3	53	1.3s	262k	$0.60	$2.5	2025-09-05
o1	OpenAI	closed	15.2	·	·	200k	$15	$60	2024-12-05
GLM-4.7-Flash	Z AI	open	14.9	63	1.5s	203k	$0.07	$0.40	2026-01-19
DeepSeek V3.1 Terminus	DeepSeek	open	14.8	·	·	131k	$0.27	$1	2025-09-22
Grok 4.20 0309 (Non-reasoning)	SpaceXAI	closed	14.6	·	·	1M	$1.3	$2.5	·
Mistral Medium 3.5	Mistral	open	14.2	165	2.3s	262k	$1.5	$7.5	2026-04-29
Mercury 2	Inception	closed	13.8	777	4.3s	128k	$0.25	$0.75	2026-02-20
DeepSeek V3.1 (Non-reasoning)	DeepSeek	open	13.7	·	·	131k	$3	$4.5	·
Qwen3 VL 235B A22B (Reasoning)	Alibaba	open	13.4	56	3.9s	131k	$0.98	$4.0	2025-09-23
DeepSeek R1 0528	DeepSeek	open	13.1	·	·	164k	$0.50	$2.1	2025-05-28
Gemini 2.5 Flash	Google	closed	13.1	218	18s	1M	$0.30	$2.5	2025-05-20
GPT-5 nanohigh	OpenAI	closed	13	159	68s	400k	$0.05	$0.40	2025-08-07
Nemotron 3.5 Lightning	NVIDIA	open	12.9	286	0.62s	262k	$0.08	$0.20	2026-08-11
GLM-4.5	Z AI	open	12.8	·	·	131k	$0.60	$2.2	·
Nemotron 3 Super	NVIDIA	open	12.8	165	1.3s	262k	$0.085	$0.40	2026-03-11
GPT-4.1	OpenAI	closed	12.7	154	0.86s	1M	$2	$8	2025-04-14
Qwen3 235B A22B 2507	Alibaba	open	12.7	73	2.8s	131k	$0.30	$3	2025-07-25
o3-mini	OpenAI	closed	12.5	223	5.4s	200k	$1.1	$4.4	2025-01-31
o1-pro	OpenAI	closed	12.4	·	·	200k	$150	$600	·
Qwen3 Coder 480B	Alibaba	open	11.9	55	3.1s	262k	$1.5	$7.5	2025-07-22
gpt-oss-120bhigh	OpenAI	open	11.6	222	0.88s	131k	$0.06	$0.40	2025-08-05
Mistral Small 4	Mistral	open	11.3	179	0.78s	262k	$0.15	$0.60	2026-03-16
GLM-4.6V	Z AI	open	11.2	80	3.0s	131k	$0.30	$0.90	2025-12-08
Qwen3 Next 80B A3B (Reasoning)	Alibaba	open	11.2	172	2.4s	262k	$0.15	$1.2	2025-09-11
Qwen3.5 9B	Alibaba	open	11.2	35	2.4s	262k	$0.10	$0.15	2026-03-02
GLM-4.5-Air	Z AI	open	11.1	79	2.4s	131k	$0.13	$0.85	2025-07-28
Granite 4.2 8B	IBM	open	11.1	66	0.90s	131k	$0.06	$0.25	2026-08-25
GPT-4.1 mini	OpenAI	closed	10.2	152	0.78s	1M	$0.40	$1.6	2025-04-14
Llama 4 Maverick	Meta	open	10	100	0.94s	1M	$0.20	$0.80	2025-04-05
Qwen3 VL 235B A22B	Alibaba	open	9.9	53	2.7s	262k	$0.40	$1.6	2025-09-23
Qwen3 30B A3B 2507	Alibaba	open	9.8	136	2.4s	41k	$0.12	$0.50	2025-07-30
Qwen3 Coder 30B A3B	Alibaba	open	9.6	80	2.6s	160k	$0.07	$0.27	2025-07-31
Mistral Large 3	Mistral	open	9.3	78	1.0s	262k	$0.50	$1.5	2025-12-02
Qwen3 Coder Next	Alibaba	open	9.2	102	1.2s	262k	$0.20	$1.5	2026-02-03
gpt-oss-20bhigh	OpenAI	open	9	166	0.81s	131k	$0.03	$0.14	2025-08-05
Mistral Medium 3	Mistral	closed	9	·	·	131k	$0.40	$2	2025-05-07
Nemotron 3 Nano	NVIDIA	open	8.9	193	1.1s	262k	$0.05	$0.20	2025-12-15
Qwen3 32B	Alibaba	open	8.6	106	2.5s	41k	$0.08	$0.28	2025-04-28
GPT-4o (Nov)	OpenAI	closed	8.4	136	1.1s	128k	$2.5	$10	2024-11-20
Qwen3 14B	Alibaba	open	8.2	63	2.7s	41k	$0.12	$0.24	2025-04-28
Llama 4 Scout	Meta	open	8.1	133	0.84s	328k	$0.10	$0.30	2025-04-05
DeepSeek R1 Distill Llama 70B	DeepSeek	open	7.9	·	·	8k	$2	$2	·
Qwen3 VL 30B A3B	Alibaba	open	7.9	113	2.2s	262k	$0.15	$0.60	2025-10-03
GPT-4.1 nano	OpenAI	closed	7.8	182	0.74s	1M	$0.10	$0.40	2025-04-14
Solar Pro 3	Upstage	closed	7.8	141	1.3s	131k	$0.15	$0.60	2026-04-06
GPT-4o (Aug)	OpenAI	closed	7.7	104	1.2s	128k	$2.5	$10	2024-08-06
Llama 3.3 70B	Meta	open	7.7	89	1.6s	131k	$0.10	$0.32	2024-12-06
Qwen2.5 72B	Alibaba	open	7.7	·	·	33k	$0.36	$0.40	2024-09-19
GLM-4.5V	Z AI	open	7.6	43	2.8s	66k	$0.60	$1.8	2025-08-11
Hermes 4 405B	Nous Research	open	7.5	43	2.4s	131k	$1	$3	2025-08-27
Qwen3 30B A3B 2507 (Non-reasoning)	Alibaba	open	7.5	141	1.9s	262k	$0.04815	$0.19	2025-07-29
GPT-4o (May)	OpenAI	closed	7.3	133	1.0s	128k	$5	$15	2024-05-13
GPT-4o (ChatGPT)	OpenAI	closed	7.2	·	·	128k	$2.5	$10	2025-02-15
Command A	Cohere	open	7	60	1.7s	256k	$2.5	$10	2025-03-13
Llama 3.1 8B	Meta	open	6.9	140	0.91s	16k	$0.02	$0.05	2024-07-23
Gemini 2.5 Flash-Lite (Non-reasoning)	Google	closed	6.7	265	0.31s	1M	$0.10	$0.40	·
GPT-4o mini	OpenAI	closed	6.7	120	0.84s	128k	$0.15	$0.60	2024-07-18
Mistral Small 3	Mistral	open	6.7	·	·	33k	$0.05	$0.08	2025-01-30
Llama 3.1 70B	Meta	open	6.6	67	1.7s	131k	$0.40	$0.40	·
Hermes 3 - Llama-3.1 70B	Nous Research	open	6	29	2.2s	131k	$0.70	$0.70	2024-08-15
Ministral 3 14B	Mistral	open	6	79	0.94s	262k	$0.20	$0.20	2025-12-02
Phi-4	Microsoft	open	5.9	41	2.6s	16k	$0.07	$0.14	2024-12-12
Reka Flash 3	Reka AI	open	5.6	·	·	66k	$0.10	$0.20	2025-03-10
GPT-3.5 Turbo	OpenAI	closed	5.5	·	·	16k	$0.50	$1.5	·
Ministral 3 8B	Mistral	open	5.5	105	0.79s	262k	$0.15	$0.15	2025-12-02
Gemma 3 27B	Google	open	4.9	·	·	131k	$0.08	$0.16	2025-03-12
Gemma 3 4B	Google	open	4.8	·	·	131k	$0.05	$0.10	2025-03-12
Ministral 3 3B	Mistral	open	4.8	201	0.72s	131k	$0.10	$0.10	2025-12-02
Gemma 3 12B	Google	open	3.8	·	·	131k	$0.05	$0.15	2025-03-12
codestral-2508	mistral	·	·	·	·	256k	$0.30	$0.90	·
command-r-08-2024	cohere	·	·	·	·	128k	$0.15	$0.60	·
command-r-plus-08-2024	cohere	·	·	·	·	128k	$2.5	$10	·
command-r7b-12-2024	cohere	·	·	·	·	128k	$0.0375	$0.15	·
ernie-4.5-vl-424b-a47b	novita	·	·	·	·	123k	$0.42	$1.3	·
Gemini 3.1 Flash Lite	gemini	·	·	·	·	1M	$0.25	$1.5	·
Gemini 3.1 Pro	gemini	·	·	·	·	1M	$2	$12	·
GLM-5.3 FlashX	glm	·	·	·	·	1M	$0.37	$1.3	·
gpt-5-pro	openai	·	·	·	·	400k	$15	$120	·
gpt-5.2-pro	openai	·	·	·	·	400k	$21	$168	·
gpt-5.4-pro	openai	·	·	·	·	1.1M	$30	$180	·
gpt-5.5-pro	openai	·	·	·	·	1.1M	$30	$180	·
Grok 4.20 Multi-Agent (0309)	xai	·	·	·	·	1M	$1.3	$2.5	·
hermes-3-llama-3.1-405b	deepinfra	·	·	·	·	131k	$1	$1	·
Hunyuan 4 Preview	tencent	·	·	·	·	1M	$0.83	$2.5	·
Hunyuan 4 Preview (Novita)	novita	·	·	·	·	1M	$0.83	$2.5	·
Poolside Laguna S 2.1	poolside	·	·	·	·	262k	$0.00	$0.00	·
Poolside Laguna XS 2.1	poolside	·	·	·	·	262k	$0.00	$0.00	·
Mercury 2.5	inception	·	·	·	·	260k	$0.20	$0.75	·
mistral-nemo	novita	·	·	·	·	60k	$0.04	$0.17	·
mistralai/Mistral-Nemo-Instruct-2407	deepinfra	·	·	·	·	·	$0.019	$0.03	·
mythomax-l2-13b	deepinfra	·	·	·	·	4k	$0.40	$0.40	·
Palmyra X5	writer	·	·	·	·	1M	$0.60	$6	·
Reka Edge	reka	·	·	·	·	16k	$0.10	$0.10	·
Seed 2.0 Code	deepinfra	·	·	·	·	256k	$0.50	$3	·
Seed 2.0 Mini	deepinfra	·	·	·	·	256k	$0.10	$0.40	·
Seed 2.0 Pro	deepinfra	·	·	·	·	256k	$0.50	$3	·
wizardlm-2-8x22b	novita	·	·	·	·	66k	$0.62	$0.62	·
Jev 1.13decisions	TypeSafe	·	·	·	·	32k	$0.042	$0.00	·

11·local catalog

25 models

Local models

Every model conifer models pull can install; on disk is the size of the Q4_K_M file.


Gemma 4 E2B IT	Google	4.6B	3.1 GB	128k	2026-03-31	reasoning · tools
Gemma 4 E4B IT	Google	7.5B	5.0 GB	128k	2026-03-31	reasoning · tools
Gemma 4 12B IT	Google	12B	7.1 GB	256k	2026-06-03	reasoning · tools
Gemma 4 31B IT	Google	31B	18.3 GB	256k	2026-03-31	reasoning · tools
Granite-Docling 258M	IBM	0.16B	0.33 GB	8k	2025-09-17	vision
Granite 4.0 Micro 3B	IBM	3.4B	2.1 GB	128k	2025-10-02	coding · reasoning · tools
Granite 4.1 3B	IBM	3.4B	2.1 GB	128k	2026-04-16	coding · reasoning · tools
Granite 4.0 H Tiny 7B A1B	IBM	6.9B · 1B active	4.2 GB	1M	2025-10-02	coding · MoE · reasoning · tools
Granite 4.0 H Small 32B A9B	IBM	32B · 9B active	19.5 GB	1M	2025-10-02	coding · MoE · reasoning · tools
LFM2.5 230M	Liquid AI	0.23B	0.15 GB	125k	2026-06-25	tools
LFM2.5 350M	Liquid AI	0.35B	0.23 GB	125k	2026-03-25	tools
LFM2-VL 1.6B	Liquid AI	1.2B	0.70 GB	125k	2025-08-12	tools · vision
LFM2.5 1.2B Instruct	Liquid AI	1.2B	0.73 GB	125k	2026-01-05	tools
LFM2.5 1.2B Thinking	Liquid AI	1.2B	0.73 GB	125k	2026-01-20	reasoning · tools
LFM2.5-VL 1.6B	Liquid AI	1.2B	0.73 GB	125k	2026-01-05	tools · vision
LFM2.5 8B A1B	Liquid AI	8.5B · 1B active	5.2 GB	125k	2026-05-28	MoE · reasoning · tools
Ministral 3 3B Instruct 2512	Mistral AI	3.4B	2.1 GB	256k	2025-12-02	tools
Mistral Small 3.2 24B Instruct 2506	Mistral AI	24B	14.3 GB	128k	2025-06-20	tools · vision
Qwen 3.5 0.8B	Alibaba	0.75B	0.53 GB	256k	2026-03-01	reasoning · tools
Qwen3-VL 4B Instruct	Alibaba	4B	2.5 GB	256k	2025-10-11	tools · vision
Qwen 3.5 4B	Alibaba	4.2B	2.7 GB	256k	2026-03-02	coding · reasoning · tools
Qwen 3.5 9B	Alibaba	9B	5.7 GB	256k	2026-02-28	coding · reasoning · tools
Qwen 3.5 27B	Alibaba	27B	16.7 GB	256k	2026-02-24	coding · reasoning · tools
Qwen 3.6 27B	Alibaba	27B	16.8 GB	256k	2026-04-22	coding · reasoning · tools
Qwen 3.8 27B	Alibaba	27B	17.1 GB	128k	2026-08-14	coding · reasoning · tools