gpt-oss-20b
OpenAI · Open weights · Released Aug 5, 2025 · openai/gpt-oss-20b
Best $0.017 in / $0.085 out via Darkbloom · fp8
131K contextToolsReasoningStructured outputPrompt cachingOpen weights
Available from 12 providersUSD / MTOK · PRICES 16 MINUTES AGO · SPEED 6 MINUTES AGO
EST. MONTHLY COSTclick a row for detail · click headers to sort
| PROVIDER | INPUT | OUTPUT | CACHE | CONTEXT | QUANT | TTFT | TPS | SMOKE | UPTIME |
|---|---|---|---|---|---|---|---|---|---|
| $0.017 | $0.085 | $0.009 | 131K | fp8 | 1.3s | 157 tok/s | — | 100.0% | |
| $0.018 | $0.09 | $0.009 | 131K | fp4 | 1.8s | 140 tok/s | — | 100.0% | |
| $0.024 | $0.112 | $0.012 | 131K | bf16 | ✕ | ✕ | 2/5 | 0.0% | |
| DEDekaLLM | $0.029 | $0.14 | $0.029 | 131K | bf16 | 2.1s | 59 tok/s | — | 38.9% |
| $0.03 | $0.13 | $0.03 | 131K | fp4 | 0.31s | 259 tok/s | 5/5 | 99.9% | |
| $0.03 | $0.14 | $0.015 | 131K | bf16 | — | — | — | 99.7% | |
| $0.03 | $0.15 | $0.02 | 131K | fp4 | 0.8s | 150 tok/s | 5/5 | 99.9% | |
| $0.04 | $0.18 | — | 131K | fp8 | 1.8s | 105 tok/s | 4/5 | 97.1% | |
| $0.07 | $0.15 | — | 131K | — | 1.3s | 137 tok/s | — | 100.0% | |
| $0.07 | $0.15 | — | 131K | — | — | — | — | 100.0% | |
| $0.075 | $0.30 | $0.037 | 131K | — | 0.84s | 601 tok/s | — | 99.7% | |
| $0.186 | $0.278 | $0.167 | 131K | fp4 | 0.73s | 113 tok/s | — | 94.5% |
TTFT/TPS: median (P50) over trailing 504h · chat10k profile (~10K in / 200 out) · measured via OpenRouter routing · ✕ = probed, endpoint returned no measurable stream · SMOKE: of 5 programmatic capability probes passed (JSON schema, instruction following, tool call, long-context, benign compliance; retry-once) · screened via OpenRouter routing · v0.2 · methodology v1.1 · measured 6 minutes ago
SERVE THIS MODEL AND NOT LISTED? LIST YOUR INFERENCE →
Input price history · best listed price, observed
↑ 325% SINCE NOV 3, 2026
$0.004 · NOV 3, 2026$0.017 · TODAY
Pricing variants · USD / MTok in / out
STANDARD
$0.017 / $0.085
per MTok in / out
BATCH
$0.024 / $0.112
async batch
CACHE READ $0.009
Recent changesRSS ↗
Nov 21, 2026price decreased · input0.018 → 0.017
Nov 21, 2026price decreased · output0.09 → 0.085
Nov 18, 2026price increased · output0.03 → 0.09
Nov 14, 2026price increased · output0.03 → 0.035
Nov 9, 2026price increased · input0.004 → 0.018
Nov 9, 2026price increased · output0.017 → 0.03
Nov 6, 2026price decreased · output0.037 → 0.017
Nov 5, 2026endpoint removed · cheaperinference12 → 11
About
gpt-oss-20b is an open-weight 21B parameter model released by OpenAI under the Apache 2.0 license. It uses a Mixture-of-Experts (MoE) architecture with 3.6B active parameters per forward pass, optimized for...
Specifications
Context window131,072
Max output117,964
Modalities intext
Modalities outtext
LicenseOpen weights
ReleasedAug 5, 2025
Knowledge cutoff2024-06-30
TokenizerGPT
WeightsHugging Face ↗
Related
Embed badge
[](https://modelindex.ai/models/openai/gpt-oss-20b)Sources
OpenRouter API16 minutes ago
AGGREGATORHugging Face16 minutes ago
COMMUNITYModelIndex benchmarks6 minutes ago
MEASUREDVerified 16 minutes ago · list prices, not negotiated rates.