The story so far
Weights, every dated article, and repo commits on one rail.
17 repo commits since the weights landed
Teal is the weights (2025-01-13), grey ticks are repo commits.
internlm3-8b-instruct
internlm Apache 2.0 out 21 months ago
8.80B parameters, text in, text out, 32K tokens of context.
Verified 2026-10-25 Rank #676 in LLMs · score 72.5 method v1 Sources 3 Downloads 30d 50.4k · #382 of 619 tracked
Do people use it?50.4kdownloads, 30 days#382 of 619 we track · below the median236 likesHugging Face2026-10-24
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.
No API provider serves it, so its downloads are people running it themselves: 47 quantized builds and 69 Spaces.
- As a service0no API provider listed
- On own hardware47quantized builds
- Built on114 fine-tunes · 5 adapters · 2 merges
- In apps69Hugging Face Spaces
Top Spaces running it: pliny-the-prompter/obliteratus ♥ 621 · internlm/internlm3-8b-instruct ♥ 7 · Kodacoda/obliteratus ♥ 2 · dumbordumber/obliteratus ♥ 2
Will it fit?
Weights only, an estimate: leave room for the KV cache. Method
Fits a 16 GB card at INT4
About 5.3 GB at INT4, 11 GB at FP8, 21 GB at BF16: weights only, so leave room for the KV cache.
Estimate params × bytes per param × 1.2 · how we estimate
What people built on it
58 derivatives on the Hub. Which build to run, by runtime: the most downloaded in each class.
58 derivatives on the Hub, most of them quantized
From the Hugging Face model tree; quantized builds in teal, fine-tunes, merges and adapters in greys.
Reference
Everything else, folded: the full tables, every source.
Benchmarks
16 rows
Benchmarks 16 rows
measured third-party leaderboard rowauthor-reported number from the model card or eval files (16)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.
What the card says
1 cited sentence
Use-case evidence
6 use cases
What it is good at 6 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 HellaSwag(10-shot) 91.2Author reports IFEval(Prompt-Strict) 79.3Author reports AlpacaEval 2.0(LC WinRate) 51.1+6 more
- Reason through hard problemsauthor-reportedAuthor reports MMLU(0-shot) 76.6Author reports MMLU-Pro(0-shot) 57.6Author reports GPQA-Diamond(0-shot) 37.4+3 more
- Write code with agentsauthor-reportedAuthor reports LiveCodeBench(2407-2409 Pass@1) 17.8Author reports HumanEval(Pass@1) 82.3
- Work across languagesauthor-reportedAuthor reports CMMLU(0-shot) 83.1
- Read long documentsauthor-reportedAuthor reports RULER(4-128K Average) 87.9
- Run on a laptop or phonestructuraldecision-model estimate p=0.90fits: single-gpu
Facts
20 fields
Listed in LLMs
Facts
- internlm3-8b-instruct has 8.80B parameters. (safetensors metadata, observed 2026-10-24)
- Context length is 32,768 tokens (32K). (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 2025-01-13, last modified 2025-02-11. (HF API, observed 2026-10-24)
| Pipeline tag | text-generation |
|---|---|
| Architectures | InternLM3ForCausalLM |
| Model type | internlm3 |
| Dtype | bfloat16 |
| Params by dtype | BF16: 8,804,241,408 |
| Size band | M |
| MoE | no |
| Context length | 32,768 (config.json max_position_embeddings) |
| Papers | arXiv:2403.17297 |
| License | Apache 2.0 (apache-2.0) |
| Hub snapshot | 2026-10-25 |
Servability
vLLM · SGLang
Servability
vLLM: supported · InternLM3ForCausalLM · 7050ef881ff0835f6890a47db1859a24c65481d6SGLang: supported · InternLM3ForCausalLM · 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.
Hardware fit
3 estimates
Hardware fit ESTIMATE
Weights-only memory for internlm3-8b-instruct: params × bytes/param × 1.2 overhead. KV cache and activations are extra. How this is computed.
| BF16 (2 bytes/param) | 21 GB · Single GPU (<24 GB) |
|---|---|
| FP8 (1 byte/param) | 11 GB · Single GPU (<24 GB) |
| INT4 (0.5 bytes/param) | 5.3 GB · Single GPU (<24 GB) |
Weights only (params × bytes/param × 1.2 overhead); KV cache and activations extra.
All variants
40 variants
Lineage
quantized: 47merge: 2finetune: 4adapter: 5
Top variants (40 shown, ranked by downloads)
| Variant | Relation | Quant | Params | Band | License | Author | Likes | Downloads |
|---|---|---|---|---|---|---|---|---|
| internlm/internlm3-8b-instruct-gguf | quantized | gguf | — | — | apache-2.0 | internlm | 28 | 25.7k |
| bartowski/internlm3-8b-instruct-GGUF | quantized | GGUF | — | — | apache-2.0 | bartowski | 3 | 24k |
| mradermacher/internlm3-8b-instruct-i1-GGUF | quantized | GGUF | — | — | apache-2.0 | mradermacher | 0 | 5.4k |
| lmstudio-community/internlm3-8b-instruct-GGUF | quantized | GGUF | — | — | apache-2.0 | lmstudio-community | 2 | 4.2k |
| second-state/internlm3-8b-instruct-GGUF | quantized | GGUF | — | — | apache-2.0 | second-state | 0 | 3.1k |
| DevQuasar-8/internlm.internlm3-8b-instruct-GGUF | quantized | GGUF | — | — | — | DevQuasar-8 | 0 | 2.2k |
| mradermacher/internlm3-8b-instruct-GGUF | quantized | GGUF | — | — | apache-2.0 | mradermacher | 1 | 2.1k |
| DuoNeural/InternLM3-8B-Instruct-GGUF | quantized | GGUF | — | — | apache-2.0 | DuoNeural | 0 | 1.9k |
| gaianet/internlm3-8b-instruct-GGUF | quantized | GGUF | — | — | apache-2.0 | gaianet | 0 | 1.5k |
| AbteeXAILab/lumynax-reasoning-internlm3-8b-gguf | quantized | gguf | — | — | apache-2.0 | AbteeXAILab | 0 | 902 |
| mlx-community/internlm3-8b-instruct-4bit | quantized | -4bit | 8.80B | M | apache-2.0 | mlx-community | 3 | 634 |
| OpenGVLab/InternVL3-9B-Pretrained | merge | — | 9.14B | L | mit | OpenGVLab | 0 | 765 |
| silence09/InternLM3-8B-Instruct-Converted-LlaMA | finetune | — | 8.80B | M | apache-2.0 | silence09 | 6 | 336 |
| silence09/InternLM3-8B-Instruct-Converted-Qwen2 | finetune | — | 8.80B | M | apache-2.0 | silence09 | 3 | 235 |
| ToastyPigeon/intern-rp-lora | adapter | GGUF | — | — | apache-2.0 | ToastyPigeon | 0 | 310 |
| roleplaiapp/internlm3-8b-instruct-Q4_K_M-GGUF | quantized | GGUF | — | — | — | roleplaiapp | 1 | 227 |
| jakiAJK/internlm3-8b-instruct_AWQ | quantized | AWQ | 8.80B | M | — | jakiAJK | 0 | 252 |
| allura-org/ILM3-8B-Ruby-Music | finetune | — | 8.80B | M | apache-2.0 | allura-org | 0 | 214 |
| roleplaiapp/internlm3-8b-instruct-Q3_K_S-GGUF | quantized | GGUF | — | — | — | roleplaiapp | 0 | 212 |
| jakiAJK/internlm3-8b-instruct_GPTQ-int4 | quantized | GPTQ | 8.80B | M | — | jakiAJK | 0 | 204 |
| roleplaiapp/internlm3-8b-instruct-Q2_K-GGUF | quantized | Q2_K | — | — | — | roleplaiapp | 0 | 199 |
| roleplaiapp/internlm3-8b-instruct-Q3_K_L-GGUF | quantized | GGUF | — | — | — | roleplaiapp | 0 | 197 |
| fbaldassarri/internlm_internlm3-8b-instruct-autoround-int8-gs64-asym | quantized | int8 | 3.14B | M | apache-2.0 | fbaldassarri | 0 | 193 |
| roleplaiapp/internlm3-8b-instruct-Q3_K_M-GGUF | quantized | GGUF | — | — | — | roleplaiapp | 0 | 192 |
| roleplaiapp/internlm3-8b-instruct-Q4_0-GGUF | quantized | Q4_0 | — | — | — | roleplaiapp | 0 | 192 |
| roleplaiapp/internlm3-8b-instruct-Q4_K_S-GGUF | quantized | GGUF | — | — | — | roleplaiapp | 0 | 190 |
| roleplaiapp/internlm3-8b-instruct-Q5_0-GGUF | quantized | Q5_0 | — | — | — | roleplaiapp | 0 | 172 |
| roleplaiapp/internlm3-8b-instruct-Q6_K-GGUF | quantized | Q6_K | — | — | — | roleplaiapp | 0 | 171 |
| roleplaiapp/internlm3-8b-instruct-Q5_K_M-GGUF | quantized | GGUF | — | — | — | roleplaiapp | 0 | 167 |
| roleplaiapp/internlm3-8b-instruct-Q5_K_S-GGUF | quantized | GGUF | — | — | — | roleplaiapp | 0 | 162 |
| roleplaiapp/internlm3-8b-instruct-Q8_0-GGUF | quantized | Q8_0 | — | — | — | roleplaiapp | 0 | 137 |
| FriendliAI/InternVL3-9B | merge | — | 9.14B | L | mit | FriendliAI | 0 | 130 |
| fbaldassarri/internlm_internlm3-8b-instruct-autogptq-int4-gs64-asym | quantized | gptq | 8.80B | M | apache-2.0 | fbaldassarri | 0 | 125 |
| fbaldassarri/internlm_internlm3-8b-instruct-autoround-int8-gs128-sym | quantized | int8 | 3.07B | M | apache-2.0 | fbaldassarri | 0 | 122 |
| nqzfaizal77ai/tian-zhi-reinit-550m-zero | finetune | — | 551M | S | — | nqzfaizal77ai | 0 | 110 |
| ctranslate2-4you/InternLM3-8b-Instruct-ct2-int8 | quantized | int8 | — | — | apache-2.0 | ctranslate2-4you | 0 | 109 |
| fbaldassarri/internlm_internlm3-8b-instruct-autogptq-int4-gs128-asym | quantized | gptq | 8.80B | M | apache-2.0 | fbaldassarri | 0 | 107 |
| fbaldassarri/internlm_internlm3-8b-instruct-autogptq-int8-gs128-sym | quantized | gptq | 8.80B | M | apache-2.0 | fbaldassarri | 0 | 99 |
| ctranslate2-4you/InternLM3-8b-Instruct-ct2-AWQ | quantized | AWQ | — | — | apache-2.0 | ctranslate2-4you | 0 | 96 |
| fbaldassarri/internlm_internlm3-8b-instruct-autoround-int4-gs128-asym | quantized | int4 | 2.09B | M | apache-2.0 | fbaldassarri | 0 | 90 |
Repository changelog
17 commits
Repository changelog 17 commits
| Date | Commit | SHA |
|---|---|---|
| 2025-02-11 | Update modeling_internlm3.py (#18) | 28c9941 |
| 2025-01-16 | Upload README.md (#15) | 011ff6d |
| 2025-01-16 | update readme (#11) | dc29aeb |
| 2025-01-16 | Update the installation of vllm in readme (#13) | 32ab132 |
| 2025-01-16 | Update README.md (#14) | 8955cb2 |
| 2025-01-15 | update readme (#6) | 4b3bebb |
| 2025-01-15 | use bf16 ckpt to fp32 ckpt | efa91a9 |
| 2025-01-15 | Add License as mentioned in model card (#5) | 6742f2e |
| 2025-01-15 | upload new ckpt | 7dc978d |
| 2025-01-15 | Delete model-00002-of-00004.safetensors | 8566460 |
| 2025-01-15 | Delete model-00001-of-00004.safetensors | 009201e |
| 2025-01-15 | Delete model-00004-of-00004.safetensors | 38623fc |
| 2025-01-15 | Delete model-00003-of-00004.safetensors | a939f29 |
| 2025-01-15 | Upload README.md (#2) | 4e13d8e |
| 2025-01-15 | Upload README.md (#1) | 5e4eefa |
| 2025-01-13 | update model | 790e9de |
| 2025-01-13 | initial commit | c7908a8 |
Papers
1 paper
Papers 1
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).
- InternLM2 Technical Report this model
Similar models
3 models
Similar models
- Llama-3.1-8B-Instruct · 8.03B same size band (M), same task (text-generation), more providers
- Qwen3-8B · 8.19B same size band (M), same task (text-generation), more providers, permissive license
- Meta-Llama-3-8B-Instruct · 8.03B same size band (M), same task (text-generation), more providers
Sources
3 sources
Sources
Hub snapshot: 2026-10-25 · HF API: 2026-10-24 21:42 UTC · Providers: 2026-10-24 21:44 UTC · Articles: 2026-10-24 21:46 UTC · Summary: 2026-10-07 15:15 UTC · Last verified 2026-10-25
Follow internlm3-8b-instruct
Watch this model
- /feeds/models/internlm--internlm3-8b-instruct.xml(RSS 2.0) · every change to internlm3-8b-instruct
- /feeds/families/internlm--internlm3.xml(RSS 2.0) · every change across the internlm3 family
Email alerts are not open yet; the feeds above are the same data. See feeds for the item schema.