Quality
Known Good Index and the capability indices, side by side
Each index is min-max normalised across tracked models and scaled 0–100 from published evaluations. We do not run these evaluations.
| Metric | OpenAI: gpt-oss-120b | Z.ai: GLM 4.7 |
|---|---|---|
| Known Good Index | 61 | 68 |
| Mathematics Index | 89 | 83 |
| Agentic Index | 18 | 35 |
| Openness Index | 60 | 60 |
Quality Breakdown
Every evaluation both models report a published score for
Bar = OpenAI: gpt-oss-120b · marker = Z.ai: GLM 4.7 · scores ingested from Epoch AI (CC-BY 4.0)
Right-hand figures read OpenAI: gpt-oss-120b / Z.ai: GLM 4.7. OpenAI: gpt-oss-120b leads on 1 of 3, Z.ai: GLM 4.7 on 2.
Arena Elo
Human preference rating from blind pairwise votes
Elo from blind pairwise votes, ingested from Arena. Higher is better.
| Metric | OpenAI: gpt-oss-120b | Z.ai: GLM 4.7 |
|---|---|---|
| Text Arena | 1,366 | 1,436 |
| Text Arena (style controlled) | 1,352 | 1,442 |
Price & Cost
Published API pricing by token type
USD per 1M tokens. Blended is a 3:1 input:output mix. Live from OpenRouter, cross-checked against models.dev. Lower is better.
| Metric | OpenAI: gpt-oss-120b | Z.ai: GLM 4.7 |
|---|---|---|
| Input $/1M | $0.04 | $0.40 |
| Output $/1M | $0.17 | $1.75 |
| Blended $/1M | $0.07 | $0.74 |
| Cache read $/1M | $0.015 | $0.080 |
Context Window
Maximum input accepted and maximum output emitted
Provider-declared limits, from metadata via models.dev and OpenRouter. Higher is better.
| Metric | OpenAI: gpt-oss-120b | Z.ai: GLM 4.7 |
|---|---|---|
| Context window | 131k | 205k |
| Max output tokens | 131k | 131k |
Speed & Latency
Median throughput and time to first token across hosting providers
Medians across all tracked providers, derived from OpenRouter endpoint telemetry.
| Metric | OpenAI: gpt-oss-120b | Z.ai: GLM 4.7 |
|---|---|---|
| Output tokens/s | 145 | 60 |
| Time to first token | 0.9s | 3.1s |
Providers
Which hosts serve each model, and at what price and speed
| Provider | OpenAI: gpt-oss-120b blended | OpenAI: gpt-oss-120b tok/s | Z.ai: GLM 4.7 blended | Z.ai: GLM 4.7 tok/s |
|---|---|---|---|---|
| Cerebras | $0.45 | 501 | $2.38 | 310 |
| SambaNova | $0.34 | 367 | — | — |
| Groq | $0.26 | 352 | — | — |
| Amazon Bedrock · first party | $0.26 | 342 | — | — |
| Nebius | $0.26 | 231 | — | — |
| DeepInfra | $0.07 | 196 | $0.74 | 38 |
| Google · first party | $0.16 | 177 | $1.00 | 122 |
| BaseTen | $0.20 | 171 | — | — |
| Mara | $0.30 | 146 | — | — |
| Together | $0.26 | 114 | — | — |
| Parasail | $0.26 | 101 | — | — |
| Phala | $0.26 | 82 | $1.46 | 36 |
| Novita | $0.10 | 64 | $0.90 | 21 |
| Mancer 2 | $0.17 | 40 | — | — |
| CoreWeave | $0.07 | 32 | — | — |
| DigitalOcean | $0.17 | 22 | — | — |
| SiliconFlow | $0.15 | 12 | — | — |
| AtlasCloud | — | — | $0.85 | 35 |
| Venice | — | — | $1.07 | 33 |
| StreamLake | — | — | $0.80 | 24 |
| Z.AI · first party | — | — | $1.00 | 21 |
Per-host pricing and telemetry from OpenRouter. A dash means that host does not serve that model.