openchat_3.5

Text in, text out, 8K tokens of context. 1 provider serves it.

Verified 2026-10-25 Rank #487 in LLMs · score 76.9 method v1 Sources 3 Downloads 30d 715 · #602 of 619 tracked

Can I use it?Apache 2.0Permissive; patent grantlicence file2026-10-24

Do people use it?715downloads, 30 days#602 of 619 we track · below the median+2 likes, downloads +4% this week1,145 likesHugging Face2026-10-24

Worth reading

Picked by rule from 3 tracked pieces: the lab's own post, the most shared, a field report, the latest.

Latest · 3 years ago

OpenChat-3.5 Update 0106

twitter.com 2024-01-10 1 HN pts

3 best of 3 tracked

The 1 pick above, plus these 2. Top is the radar's score: discussion, source authority and the lab's own posts, fading with age.

DiscussionOpenChat-3.5-1210: Best Open Source 7B LLMtwitter.com · 2023-12-22 · 3 HN pts
DiscussionOpen Chat 3.5old.reddit.com · 2023-11-03 · 1 HN pts

Where to run it

Live hosted endpoints, cheapest first, USD per million tokens. Ties break on output price, then provider name.

#	Cheapest live row	In /1M	Out /1M	Context	Quant	Seen
1	Featherless	—	—	—	—	Oct 25

All 1 provider →

Where it's used

Hugging Face does not say who downloads a model or with what. These are the proxies: API providers listing it, quantized builds, fine-tunes/adapters/merges, and Spaces that load it.

Served by 1 API provider; 13 quantized builds for local use; a big base for builders (52 fine-tunes, adapters and merges).

As a service1Featherless
On own hardware13quantized builds
Built on5228 fine-tunes · 17 adapters · 7 merges
In apps100+Hugging Face Spaces

Top Spaces running it: radinhas/hf-llm-api ♥ 14 · NiansuhAI/Main ♥ 10 · sepioo/facebook-translation ♥ 10 · CATIE-AQ/Guide_Evaluation_LLM ♥ 5

The story so far

Weights, every dated article, and repo commits on one rail.

3 dated articles and 30 repo commits so far

Teal is the weights (2023-10-30), amber triangles are 3 dated articles (each links to it), grey ticks are repo commits.

What people built on it

65 derivatives on the Hub. Which build to run, by runtime: the most downloaded in each class.

65 derivatives on the Hub, led by 28 fine-tunes

From the Hugging Face model tree; quantized builds in teal, fine-tunes, merges and adapters in greys.

GGUF TheBloke/openchat_3.5-GGUF for llama.cpp, Ollama, LM Studio 84.6k downloads GPTQ TheBloke/openchat_3.5-GPTQ for vLLM, SGLang 51.8k downloads

All variants →

Reference

Everything else, folded: the full tables, every source.

Benchmarks

14 rows

Benchmarks 14 rows

measured third-party leaderboard rowauthor-reported number from the model card or eval files (14)No confidence intervals unless the source gave them; settings differ between rows, so compare within a benchmark, not across. Date = the benchmark's own date when it gives one, else when we read it.

Benchmark	Value	Setting / notes	Reported by	Date
author-reported AGIEval	47.4	OpenChat-3.5	model author	2026-10-07
author-reported Average	56.4	OpenChat 3.5	model author	2026-10-07
author-reported Average	61.6	OpenChat-3.5	model author	2026-10-07
author-reported BBH CoT	63.5	OpenChat-3.5	model author	2026-10-07
author-reported BBH MC	47.6	OpenChat-3.5	model author	2026-10-07
author-reported GSM8K · Reason through hard problems	77.3	OpenChat 3.5	model author	2026-10-07
author-reported GSM8K · Reason through hard problems	77.3	OpenChat-3.5	model author	2026-10-07
author-reported HumanEval · Write code with agents	55.5	OpenChat 3.5	model author	2026-10-07
author-reported HumanEval · Write code with agents	55.5	OpenChat-3.5	model author	2026-10-07
author-reported MATH-500 · Reason through hard problems	28.6	OpenChat 3.5	model author	2026-10-07
author-reported MMLU · Reason through hard problems	64.3	OpenChat 3.5	model author	2026-10-07
author-reported MMLU · Reason through hard problems	64.3	OpenChat-3.5	model author	2026-10-07
author-reported MT-Bench · Chat and follow instructions	7.81	OpenChat-3.5	model author	2026-10-07
author-reported TruthfulQA · Chat and follow instructions	59.1	OpenChat-3.5	model author	2026-10-07

Use-case evidence

4 use cases

What it is good at 4 use cases

Labels are evidence, not verdicts: measured a third-party leaderboard row for this model; author-reported a number the model author published in the card or eval files; structural a deterministic signal from the name, config or chat template, or a decision-model estimate. Each chip links to its source. All use cases.

Chat and follow instructionsauthor-reportedAuthor reports MT-Bench 7.81 Author reports TruthfulQA 59.1 decision-model estimate p=1.00 +2 more
Reason through hard problemsauthor-reportedAuthor reports MMLU 64.3 Author reports MATH 28.6 Author reports GSM8k 77.3
Write code with agentsauthor-reportedAuthor reports HumanEval 55.5
Run on a laptop or phonestructuraldecision-model estimate p=0.98

Facts

18 fields

Listed in LLMs

Facts

Context length is 8,192 tokens (8K). (config.json max_position_embeddings, observed 2026-10-24)
License: Apache 2.0 (friction: none). Permissive; patent grant. (license file, observed 2026-10-24)
Created 2023-10-30, last modified 2024-05-18. (HF API, observed 2026-10-24)

Pipeline tag	text-generation
Library	transformers
Architectures	MistralForCausalLM
Model type	mistral
Dtype	bfloat16
MoE	no
Context length	8,192 (config.json max_position_embeddings)
Papers	arXiv:2309.11235, arXiv:2303.08774
License	Apache 2.0 (apache-2.0)
Hub snapshot	2026-10-25

Servability

vLLM · SGLang

Servability

vLLM: supported · MistralForCausalLM · 7050ef881ff0835f6890a47db1859a24c65481d6SGLang: supported · MistralForCausalLM · 773cb938fe9484a8f7e7e46fb1b1fe96412c59ed

vLLM registry (7050ef881ff0835f6890a47db1859a24c65481d6) · SGLang registry (773cb938fe9484a8f7e7e46fb1b1fe96412c59ed) · Matched by architecture class name against each project's registry on main; 'supported' means the architecture is registered, not that this checkpoint was tested.

All variants

40 variants

Lineage

Original model Family: openchat/openchat-3.5 65 derivatives on the Hub

quantized: 13finetune: 28merge: 7adapter: 17

Top variants (40 shown, ranked by downloads)

Variant	Relation	Quant	Params	Band	License	Author	Likes	Downloads
TheBloke/openchat_3.5-GGUF	quantized	GGUF	—	—	apache-2.0	TheBloke	133	84.6k
TheBloke/openchat_3.5-GPTQ	quantized	GPTQ	7.24B	M	apache-2.0	TheBloke	17	51.8k
FuseAI/FuseChat-7B-VaRM	finetune	—	7.24B	M	apache-2.0	FuseAI	94	20.6k
TheBloke/openchat_3.5-AWQ	quantized	AWQ	7.24B	M	apache-2.0	TheBloke	15	24.2k
AIDX-ktds/ktdsbaseLM-v0.12-based-on-openchat3.5	finetune	—	10.8B	L	apache-2.0	AIDX-ktds	8	19.1k
mradermacher/openchat_3.5-i1-GGUF	quantized	GGUF	—	—	apache-2.0	mradermacher	1	13.5k
second-state/OpenChat-3.5-GGUF	quantized	GGUF	—	—	apache-2.0	second-state	3	12.3k
rishiraj/smol-7b	finetune	—	7.24B	M	apache-2.0	rishiraj	23	9.8k
FuseAI/OpenChat-3.5-7B-Solar	finetune	—	7.24B	M	apache-2.0	FuseAI	9	7.8k
FuseAI/OpenChat-3.5-7B-Mixtral	finetune	—	7.24B	M	apache-2.0	FuseAI	10	7.3k
nlpguy/StockFuseChat	merge	—	7.24B	M	apache-2.0	nlpguy	1	7.3k
FuseAI/FuseChat-7B-Slerp	finetune	—	7.24B	M	apache-2.0	FuseAI	6	6.8k
mradermacher/openchat_3.5-GGUF	quantized	GGUF	—	—	apache-2.0	mradermacher	0	6.8k
FuseAI/FuseChat-7B-TA	finetune	—	7.24B	M	apache-2.0	FuseAI	6	6.2k
nlpguy/StarFusion-alpha1	merge	—	7.24B	M	apache-2.0	nlpguy	1	5k
nlpguy/StarFusion-alpha2	merge	—	7.24B	M	apache-2.0	nlpguy	1	4.9k
tensorblock/openchat_3.5-GGUF	quantized	GGUF	—	—	apache-2.0	tensorblock	0	2k
FuseAI/FuseChat-7B-v2.0	finetune	—	7.24B	M	apache-2.0	FuseAI	10	1.2k
FuseAI/OpenChat-3.5-7B-Starling-v2.0	finetune	—	7.24B	M	apache-2.0	FuseAI	2	863
FuseAI/OpenChat-3.5-7B-InternLM-v2.0	finetune	—	7.24B	M	apache-2.0	FuseAI	1	851
FuseAI/OpenChat-3.5-7B-Qwen-v2.0	finetune	—	7.24B	M	apache-2.0	FuseAI	0	881
FuseAI/OpenChat-3.5-7B-SOLAR-v2.0	finetune	—	7.24B	M	apache-2.0	FuseAI	1	800
FuseAI/OpenChat-3.5-7B-Mixtral-v2.0	finetune	—	7.24B	M	apache-2.0	FuseAI	0	823
ysn-rfd/openchat_3.5-Q8_0-GGUF	quantized	Q8_0	—	—	apache-2.0	ysn-rfd	0	748
neboolaai/NAI-3.5	quantized	—	7.24B	M	mit	neboolaai	5	395
Cornell-AGI/REBEL-OpenChat-3.5	finetune	—	—	—	apache-2.0	Cornell-AGI	1	500
nlpguy/StarFusion-alpha2-Q4_K_M-GGUF	merge	GGUF	—	—	apache-2.0	nlpguy	0	340
SEOKDONG/openchat3.5_korean_v1.0_sft	finetune	—	7.24B	M	apache-2.0	SEOKDONG	3	182
brekJ/openchat_3.5-Q4_K_M-GGUF	quantized	GGUF	—	—	apache-2.0	brekJ	0	321
itlwas/openchat_3.5-Q4_K_M-GGUF	quantized	GGUF	—	—	apache-2.0	itlwas	0	302
Dr-Floofers/PIPv2	adapter	—	—	—	—	Dr-Floofers	0	268
zuhashaik/MELD-EFR-Z	adapter	—	—	—	mit	zuhashaik	1	205
ysn-rfd/openchat_3.5-GGUF	quantized	GGUF	—	—	apache-2.0	ysn-rfd	0	223
LoneStriker/smol-7b-6.0bpw-h6-exl2	finetune	0bpw	—	—	apache-2.0	LoneStriker	0	212
pmarkovic/openchat-35-sft	finetune	—	7.24B	M	apache-2.0	pmarkovic	0	199
LoneStriker/smol-7b-5.0bpw-h6-exl2	finetune	0bpw	—	—	apache-2.0	LoneStriker	0	189
baebee/7B-Cetacea	merge	—	7.24B	M	—	baebee	0	189
SHIBA212121/openchat_3.5-Q5_K_M-GGUF	quantized	GGUF	—	—	apache-2.0	SHIBA212121	0	182
AiJoker/openchat_3.5-slp-jordan-belfort	finetune	—	7.24B	M	—	AiJoker	0	139
amgadhasan/scachase-openchat-0.2	finetune	—	7.24B	M	apache-2.0	amgadhasan	0	88

Repository changelog

30 commits

Repository changelog 30 commits

Date	Commit	SHA
2024-05-18	Update README.md	0fc98e3
2024-05-07	Update README.md	c8ac815
2023-12-24	Update README.md	e4d87b9
2023-12-13	Update README.md	99d59d4
2023-12-08	Update README.md	2b2dbf0
2023-12-08	Update README.md	effe26a
2023-11-23	Update README.md	75a0c61
2023-11-22	Update README.md	7e65595
2023-11-21	Add chat template (#27)	a11af6b
2023-11-17	Create generation_config.json (#21)	5b874a3
2023-11-15	Update README.md	600a344
2023-11-09	Update model card metadata	2d39b2e
2023-11-08	Update dataset details in model card (#11)	37ca7ce
2023-11-05	Update README.md	0be788e
2023-11-05	Update README.md	b6c8863
2023-11-05	Update README.md	5563fa7
2023-11-05	Update README.md	17ccd19
2023-11-04	Update README.md	eebadee
2023-11-04	Add baselines	6814591
2023-11-03	Update README.md	4d04f08
2023-11-03	Update README.md	3618d9b
2023-11-03	EOS should be 32000 (#4)	d4024c4
2023-11-03	EOS should be 32000 (#3)	394df69
2023-11-02	[doc] update README	f85ea3d
2023-11-02	Update model card	17083ef
2023-11-01	Add fast tokenizer support	81689a2
2023-10-30	bf16	0ba9dfe
2023-10-30	fp16	1387ce5
2023-10-30	weights	1025f2b
2023-10-30	initial commit	f897c6b

All providers

1 provider · 1 endpoint

Where to run 1 providers · 1 endpoints

Hosted endpoints observed for openchat_3.5, sorted by input price (USD per million tokens). Each row names its observer and date; prices change daily, so check the source before quoting.

Provider	Model ref	Context	Quant	$/M in	$/M out	tok/s	TTFT	Tools	Status	Observed
Featherless hf	openchat/openchat_3.5	—	—	—	—	—	—	—	live	2026-10-25

Articles and coverage

3 pieces

Ranked by source signals (HN, Lobsters, authority domains, official posts) and recency.

Community3

OpenChat-3.5-1210: Best Open Source 7B LLM

twitter.com2023-12-22HN 3 pts · 0 comments

Open Chat 3.5

old.reddit.com2023-11-03HN 1 pts · 0 comments

OpenChat-3.5 Update 0106

twitter.com2024-01-10HN 1 pts · 0 comments

Papers

2 papers

Papers 2

arXiv links found in the model card and its tags. "Peer-reviewed" is shown only with evidence: a published venue in OpenAlex (measured) or the authors' own journal reference on arXiv (reported).

OpenChat: Advancing Open-source Language Models with Mixed-Quality Data this model

Guan Wang, Sijie Cheng, Xianyuan Zhan +3 · 2023-09-20 · PDF · HF paper page (15 upvotes)
GPT-4 Technical Report cited in card

OpenAI, Josh Achiam, Steven Adler +278 · 2023-03-15 · PDF · HF paper page (7 upvotes)

Similar models

3 models

Similar models

opt-125m
same task (text-generation)
gpt-j-6b
same task (text-generation), permissive license
DialoGPT-medium
same task (text-generation), permissive license

Sources

3 sources

Sources

Hub snapshot: 2026-10-25 · HF API: 2026-10-24 21:40 UTC · Providers: 2026-10-24 21:44 UTC · Articles: 2026-10-24 21:46 UTC · Summary: 2026-10-07 15:12 UTC · Last verified 2026-10-25

Follow openchat_3.5

Watch this model

/feeds/models/openchat--openchat_3.5.xml(RSS 2.0) · every change to openchat_3.5
/feeds/families/openchat--openchat-3.5.xml(RSS 2.0) · every change across the openchat-3.5 family

Email alerts are not open yet; the feeds above are the same data. See feeds for the item schema.