Model providers · benchmark source
Provider catalogs
System limits from provider documentation — context capacity is not a quality benchmark.
System capacity
Context window
The amount of input and output context a model can process in one interaction under the provider's rules.
Why it matters
It sets a hard ceiling for documents, repositories, and conversation state.
Use it when
You are sizing a real workload and will test long-context quality, truncation, and cost.
The limitation
More tokens do not guarantee better retrieval, attention, or reasoning across the whole window.
Leaderboard snapshot
Context window results in a readable view
Use the tabs for the primary score and other published fields from this board. The source link remains authoritative for the live table.
Local leader
GPT-5.6 Sol
1.05M
Rows shown
25
Context tokens
Snapshot date
2026-08
14 days old · not a live API feed
Score profile
Context tokens by published row
More capacity is not automatically better
GPT-5.6 SolGPT-5.6 TerraGPT-5.6 Luna
Gemini 3.1 ProGemini 3.6 FlashGemini 3.7 FlashGemini 3.5 Flash-Lite
Kimi K3
Muse Spark 1.2
Claude Opus 5Claude Fable 5Claude Sonnet 5
GLM 5.2GLM 5.3
DeepSeek V4-FlashDeepSeek V4-Pro
Qwen3.8-Max
Grok 4.5Grok 4.6Kimi K2.6
Ling 3.0 FlashQwen3.8-2.4T-A95BQwen3.8-27BClaude Haiku 4.5Muse Glimmer 30B| Rank | Model or system | Context tokens | Other details | Note |
|---|---|---|---|---|
| #1 | GPT-5.6 SolOpenAI | 1.05M | Capacity: Provider catalog; verify the model-specific receipt | — |
| #1 | GPT-5.6 TerraOpenAI | 1.05M | Capacity: Provider catalog; verify the model-specific receipt | — |
| #1 | GPT-5.6 LunaOpenAI | 1.05M | Capacity: Provider catalog; verify the model-specific receipt | — |
| #4 | Gemini 3.1 ProGoogle | 1.05M | Capacity: Provider catalog; verify the model-specific receipt | — |
| #4 | Gemini 3.6 FlashGoogle | 1.05M | Capacity: Provider catalog; verify the model-specific receipt | — |
| #4 | Gemini 3.7 FlashGoogle | 1.05M | Capacity: Provider catalog; verify the model-specific receipt | — |
| #4 | Gemini 3.5 Flash-LiteGoogle | 1.05M | Capacity: Provider catalog; verify the model-specific receipt | — |
| #4 | Kimi K3Moonshot AI | 1.05M | Capacity: Provider catalog; verify the model-specific receipt | — |
| #4 | Muse Spark 1.2Meta | 1.05M | Capacity: Provider catalog; verify the model-specific receipt | — |
| #10 | Claude Opus 5Anthropic | 1M | Capacity: Provider catalog; verify the model-specific receipt | — |
| #10 | Claude Fable 5Anthropic | 1M | Capacity: Provider catalog; verify the model-specific receipt | — |
| #10 | Claude Sonnet 5Anthropic | 1M | Capacity: Provider catalog; verify the model-specific receipt | — |
| #10 | GLM 5.2Z.ai | 1M | Capacity: Provider catalog; verify the model-specific receipt | — |
| #10 | GLM 5.3Z.ai | 1M | Capacity: Provider catalog; verify the model-specific receipt | — |
| #10 | DeepSeek V4-FlashDeepSeek | 1M | Capacity: Provider catalog; verify the model-specific receipt | — |
| #10 | DeepSeek V4-ProDeepSeek | 1M | Capacity: Provider catalog; verify the model-specific receipt | — |
| #10 | Qwen3.8-MaxQwen | 1M | Capacity: Provider catalog; verify the model-specific receipt | — |
| #18 | Grok 4.5SpaceXAI | 500K | Capacity: Provider catalog; verify the model-specific receipt | — |
| #18 | Grok 4.6SpaceXAI | 500K | Capacity: Provider catalog; verify the model-specific receipt | — |
| #20 | Kimi K2.6Moonshot AI | 262K | Capacity: Provider catalog; verify the model-specific receipt | — |
| #20 | Ling 3.0 FlashInclusionAI | 262K | Capacity: Provider catalog; verify the model-specific receipt | — |
| #20 | Qwen3.8-2.4T-A95BQwen | 262K | Capacity: Provider catalog; verify the model-specific receipt | — |
| #20 | Qwen3.8-27BQwen | 262K | Capacity: Provider catalog; verify the model-specific receipt | — |
| #24 | Claude Haiku 4.5Anthropic | 200K | Capacity: Provider catalog; verify the model-specific receipt | — |
| #25 | Muse Glimmer 30BMeta | 131K | Capacity: Provider catalog; verify the model-specific receipt | — |
Provider-catalogued context capacities for the models in this site’s dataset. Capacity is a system limit, not a quality benchmark; the linked OpenAI documentation is an example receipt, not a source for every provider row.
Open Provider documentation example (OpenAI)Evidence in this catalog
Where Context window fits
Closest verified examples we currently carry. A missing score is not a zero.
Model examples
No directly comparable model score is verified here yet.
Harness examples
No directly comparable harness score is verified here yet.